W. Richard Bowen and Nidal Hilal 4
W. Richard Bowen and Nidal Hilal 4
W. Richard Bowen and Nidal Hilal 4
- No tags were found...
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
Butterworth-Heinemann is an imprint of Elsevier<br />
Linacre House, Jordan Hill, Oxford OX2 8DP, UK<br />
30 Corporate Drive, Suite 400, Burlington, MA 01803, USA<br />
First edition 2009<br />
Copyright © 2009, Elsevier Ltd. All rights reserved<br />
No part of this publication may be reproduced, stored in retrieval system or transmitted<br />
in any form or by any means electronic, mechanical, photocopying, recording or<br />
otherwise without the prior written permission of the publisher<br />
Permissions may be sought directly from Elsevier’s Science & Technology Rights<br />
Department in Oxford, UK: phone: (�44) (0) 1865 843830, fax: (�44) (0) 1865 853333,<br />
e-mail: permissions@elsevier.com. Alternatively you can submit your request online<br />
by visiting the Elsevier web site at http://elsevier.com/locate/permission, <strong>and</strong><br />
selecting Obtaining permission to use Elsevier material<br />
Notice<br />
No responsibility is assumed by the publisher for any injury <strong>and</strong>/or damage to<br />
persons or property as a matter of products liability, negligence or otherwise, or from<br />
any use or operation of any methods, products, instructions or ideas contained in the<br />
material herein. Because of rapid advances in the medical sciences, in particular,<br />
independent verification of diagnoses <strong>and</strong> drug dosages should be made<br />
British Library Cataloguing in Publication Data<br />
A catalogue record for this book is available from the British Library<br />
Library of Congress Cataloging in Publication Data<br />
A catalogue record for this book is available from the Library of Congress<br />
ISBN–13: 978-1-85617-517-3<br />
For information on all Butterworth-Heinemann Publications<br />
visit our web site at www.elsevierdirect.com<br />
Printed <strong>and</strong> bound in Great Britain<br />
09 10 11 12 13 10 9 8 7 6 5 4 3 2 1
Preface<br />
Atomic force microscopy (AFM) was first described in the scientific literature<br />
in 1986. It arose as a development of scanning tunnelling microscopy<br />
(STM). However, whereas STM is only capable of imaging conductive<br />
samples in vacuum, AFM has the capability of imaging surfaces at high<br />
resolution in both air <strong>and</strong> liquids. As these correspond to the conditions<br />
under which virtually all surfaces exist in the real world, this greatly<br />
increased the potentially useful role of scanning probe microscopies. This<br />
great potential of AFM led to its very rapid development. By the early<br />
1990s, it was moving outside of specialist physics laboratories <strong>and</strong> the first<br />
commercial instruments were becoming available.<br />
At the time, our main process engineering research activities were in the<br />
fields of membrane separation processes <strong>and</strong> colloid processing. Both of<br />
these fields involve the manipulation of materials on the micrometre to the<br />
nanometre length scales. To image the materials used in such processes, we<br />
used scanning electron microscopy, which was expensive, time-consuming,<br />
<strong>and</strong> even more undesirably usually involved complex sample preparation<br />
procedures <strong>and</strong> measurement in vacuum which could result in undesirable<br />
experimental artefacts. Our imagination was fired <strong>and</strong> our research greatly<br />
facilitated, following an inspiring lecture given by Jacob Israelachvili at the<br />
7th International Conference on Surface <strong>and</strong> Colloid Science in Compiègne,<br />
France, in July 1991, in which he described some of the very first applications<br />
of AFM in colloid science. Our first grant application for AFM<br />
equipment was written very shortly afterwards!<br />
Since that time there has been an enormous development of the capabilities<br />
<strong>and</strong> applicability of AFM. Physicists have devised a bewildering<br />
range of experimental techniques for probing the different properties<br />
of surfaces. Scientists, especially those working in the biological sciences,<br />
have been able to make remarkable discoveries using AFM that would<br />
have been otherwise unobtainable. A huge amount of scientific literature<br />
has appeared including a number of introductory <strong>and</strong> advanced books.<br />
However, despite the achievements <strong>and</strong> great potential for the application<br />
of AFM to process engineering, there is no book-length text describing such<br />
achievements <strong>and</strong> applications. Further, the specialist nature of the primary<br />
literature <strong>and</strong> the disciplinary strangeness of the existing book-length texts<br />
can appear rather formidable to engineers who might wish to apply AFM<br />
ix
x Preface<br />
in their work. Hence, it is our assessment that the benefit of AFM to<br />
the development of process engineering is under-fulfilled. Nevertheless, the<br />
significant decrease in cost of commercial AFM equipment, <strong>and</strong> its increasing<br />
‘user-friendliness’, has made the technique readily accessible to most<br />
engineers. We were, therefore, motivated to put together the present text<br />
with the specific intention of describing the achievements <strong>and</strong> possibilities<br />
of AFM in a way which is directly relevant to the work of our process<br />
engineering colleagues, with the hope that we will inspire them to apply<br />
this remarkable technique for the benefit of their own activities.<br />
We begin in Chapter 1 by providing an outline of the basic principles<br />
of AFM. The chapter introduces the main features of AFM equipment<br />
<strong>and</strong> describes the imaging modes which are most likely to be of benefit in<br />
process engineering applications. Such knowledge of the main operating<br />
modes should allow the reader to interpret the nature of the many subtle<br />
variations described in the primary research literature. We also introduce<br />
a remarkable benefit of AFM equipment, because it is a force microscope it<br />
can be used to directly measure surface interactions with very high resolution<br />
in both force <strong>and</strong> distance. An especially useful application of this<br />
capability is the use of ‘colloid probes’, the nature of which is introduced<br />
<strong>and</strong> the benefits of which become apparent in several of the later chapters.<br />
AFM can generate beautiful images of surfaces at subnanometre resolution.<br />
However, the detailed interpretation of the features of such images<br />
can benefit greatly from an underst<strong>and</strong>ing of the fundamental interactions<br />
from which they arise. This is the subject of Chapter 2. Depending on the<br />
materials being investigated <strong>and</strong> the experimental conditions, the interactions<br />
which give rise to such images, either separately or simultaneously,<br />
include van der Waals forces, electrical double layer forces, hydrophobic<br />
interactions, solvation forces, steric interactions, hydrodynamic drag forces<br />
<strong>and</strong> adhesion. AFM also has the capability to quantify such interactions,<br />
especially using colloid probe techniques. For this reason, mathematical<br />
descriptions of such interactions are given in forms which have proved<br />
of practical use in process engineering.<br />
Once the basics of AFM have been outlined, it is possible to move to<br />
a description of specific applications. Process engineering is a diverse<br />
<strong>and</strong> growing field comprising both established processes of great societal<br />
significance <strong>and</strong> new areas of huge promise. We begin in Chapter 3<br />
by describing investigations of an established <strong>and</strong> important type of<br />
phenomenon – the quantification of particle–bubble interactions. Such<br />
interactions are of fundamental significance in some of the largest-scale<br />
industrial processes, most notably in mineral processing <strong>and</strong> in wastewater<br />
treatment. It is especially the capability of AFM equipment to quantify the<br />
interactions between bubbles <strong>and</strong> micrometre size particles that can lead to<br />
the development of processes of increased flotation efficiency <strong>and</strong> greater<br />
specificity of separation. This is a remarkable example of how nanoscale<br />
interactions control the efficiency of megascale processes.
Preface xi<br />
Membrane separation processes are one of the most significant developments<br />
in process engineering in recent times. They now find widespread<br />
application in fields as diverse as water treatment, pharmaceutical processing,<br />
food processing, biotechnology, sensors <strong>and</strong> batteries. Membranes<br />
are most usually thin polymeric sheets, having pores in the range from<br />
the micrometre to subnanometre, that act as advanced filtration materials.<br />
Their separation capabilities are due to steric effects <strong>and</strong> the whole range<br />
of interactions that can be probed by AFM. Hence, there is a very close<br />
match between the factors that control the effectiveness of a membrane<br />
process <strong>and</strong> the measurement capabilities of AFM. In Chapter 4, we provide<br />
a survey of the numerous ways in which AFM can be used to study<br />
the factors controlling membrane processes. We consider both advanced<br />
imaging <strong>and</strong> force measurement techniques, <strong>and</strong> how they may be combined,<br />
for example, to provide a ‘visualisation’ of the rejection of a colloid<br />
particle by a membrane pore. Chapter 5 is more especially concerned with<br />
the use of AFM in the development of new membranes with specifically<br />
desirable properties. We focus, in particular, on the development of fouling<br />
resistant membranes, i.e. membranes with the minimum of unwanted<br />
adhesion of substances from the fluids being processed.<br />
In the pharmaceutical industry, there is an increasing drive to develop<br />
new ways of drug delivery, both means for the presentation of drugs to<br />
the patient <strong>and</strong> of drug formulations which target specific sites in the<br />
body. Both of these goals can benefit from knowledge of structures <strong>and</strong><br />
interactions at the nanoscale. Thus, pharmaceutical development can<br />
benefit from both the imaging <strong>and</strong> force measurement capabilities of<br />
AFM, as described in Chapter 6. The colloid probe, or more precisely<br />
drug particle probe, techniques are again very important in this work.<br />
However, there is also scope for the use of advanced techniques, such<br />
as micro- <strong>and</strong> nanothermal characterisation using a scanning thermal<br />
microscope (SThM), which can provide spatial information at a resolution<br />
unavailable to conventional calorimetry.<br />
Bioprocessing is acquiring a sophistication that was unimaginable even<br />
a few years ago. An important example is given in Chapter 7. Cells sense<br />
<strong>and</strong> respond to their surrounding microenvironment. The chapter reviews<br />
the application of micro/nanoengineering <strong>and</strong> AFM to the investigation<br />
of cell response in engineered microenvironments that mimic the natural<br />
extracellular matrix. In particular, the chapter reports the use of micro/<br />
nanoengineering to make structures that aid the underst<strong>and</strong>ing of fundamental<br />
cellular interactions, which in turn help further development of<br />
new therapeutic methods. Specific attention is given to the combination<br />
of AFM with optical microscopy for the simultaneous interrogation of<br />
biophysical <strong>and</strong> biochemical cellular processes <strong>and</strong> properties, as well as<br />
the quantification of cell viscoelasticity.<br />
Throughout the process industries, <strong>and</strong> more generally in manufacturing,<br />
the surfaces of materials are modified with coatings to protect
xii Preface<br />
them from hostile conditions <strong>and</strong> to functionalise them for a variety of<br />
purposes. In particular, ultrathin coatings play a crucial role in many<br />
processes, ranging from protection against chemical corrosion to microfabrication<br />
for microelectronics <strong>and</strong> biomedical devices. Chapter 8<br />
describes the use of AFM for the study of the fine structure <strong>and</strong> local<br />
nanomechanical properties of such advanced polymer monolayers <strong>and</strong><br />
submonolayers. AFM allows the real-time/real-space monitoring of relevant<br />
physicochemical surface processes. As miniaturisation of electronic<br />
<strong>and</strong> medical devices approaches the nanometre scale, AFM is becoming<br />
the most important characterisation tool of their nanostructural <strong>and</strong><br />
nanomechanical properties.<br />
AFM has been considered primarily as a technique for the investigation<br />
of the surfaces of solid materials, with the considerable benefit that<br />
it can be used to carry out such investigations in liquid environments.<br />
However, AFM may also be used to study the properties of such liquids<br />
themselves. This is the topic of Chapter 9, which describes dynamic studies<br />
of confined fluids, micro- <strong>and</strong> nanorheology, cavitation <strong>and</strong> adhesive<br />
failure in thin films, <strong>and</strong> meso-scale experimental studies of the tensile<br />
behaviour of thin fluid films. Such studies benefit considerably from the<br />
coupling of AFM with high-magnification optical microscopy <strong>and</strong> highspeed<br />
video techniques. The development of such studies may be of considerable<br />
importance for the many large-scale processes that depend on<br />
the properties of thin liquid films, <strong>and</strong> also for instances where the available<br />
quantities of fluids are tiny, such as for synovial fluid.<br />
In the final chapter, we have pooled the thoughts of the contributors<br />
to provide a vision of some of the ways in which AFM may contribute to<br />
the development of process engineering in the future.<br />
We thank all of the authors who have collaborated in the writing of<br />
this volume. We are very grateful for their willingness to devote time to<br />
this task <strong>and</strong> for their timely delivery of high-quality manuscripts. We<br />
also thank the many colleagues <strong>and</strong> research students who have contributed<br />
to the work described. Particular thanks are due to Dr Peter<br />
M. Williams. Peter worked as a research technician at our centre when<br />
we first started AFM studies. The results of our research as presented in<br />
this volume owe much to his technical ingenuity <strong>and</strong> patience.<br />
W. <strong>Richard</strong> <strong>Bowen</strong> <strong>and</strong> <strong>Nidal</strong> <strong>Hilal</strong><br />
wrichardbowen@i-newtonwales.org.uk<br />
nidal.hilal@nottingham.ac.uk<br />
Wales <strong>and</strong> Engl<strong>and</strong><br />
February 2009
About the Editors<br />
Professor W. <strong>Richard</strong> <strong>Bowen</strong> is a Fellow of the UK Royal Academy of<br />
Engineering. His work in chemical <strong>and</strong> biochemical engineering is widely<br />
recognised as world leading, particularly in the application of atomic force<br />
microscopy <strong>and</strong> in the development of membrane processes. He holds<br />
chairs in the Schools of Engineering at the University of Wales Swansea<br />
<strong>and</strong> the University of Surrey. He has carried out extensive consultancy for<br />
industry, government departments, research councils <strong>and</strong> universities on<br />
an international basis, currently through i-NewtonWales.<br />
xiii
xiv About thE Editors<br />
Professor <strong>Nidal</strong> <strong>Hilal</strong> is a Fellow of the Institution of Chemical Engineers<br />
<strong>and</strong> currently the Director of the Centre for Clean Water Technologies<br />
at the University of Nottingham. He obtained a PhD in Chemical<br />
Engineering from the University of Wales in 1988. Over the years, he has<br />
made a major contribution becoming an internationally leading expert in<br />
the application of Atomic Force Microscopy in process engineering <strong>and</strong><br />
membrane technology. Professor <strong>Hilal</strong> is the author of over 300 refereed<br />
publications, including 4 textbooks <strong>and</strong> 11 invited chapters in international<br />
h<strong>and</strong>books. In recognition of his substantial <strong>and</strong> sustained contribution<br />
to scientific knowledge, he was awarded a senior doctorate, Doctor<br />
of Science (DSc), from the University of Wales <strong>and</strong> the Kuwait Prize for<br />
Water Resources Development in 2005. He is a member of the editorial<br />
boards for a number of international journals <strong>and</strong> an advisor for international<br />
organizations including the Lifeboat Foundation. He is also on the<br />
panel of referees for the UK <strong>and</strong> international Research Councils.<br />
Professor <strong>Hilal</strong> acknowledges His Majesty King Abdullah Bin Abdul<br />
Aziz Al-Saud of Saudi Arabia, who is a keen advocate for nanotechnology<br />
<strong>and</strong> process engineering, particularly in the field of desalination <strong>and</strong><br />
water research for the benefit of all humanity.
List of Contributors<br />
Prof. W. <strong>Richard</strong> <strong>Bowen</strong>, FREng<br />
i-NewtonWales, 54 Llwyn y mor, Caswell, Swansea, SA3 4RD, UK<br />
wrichardbowen@i-newtonwales.org.uk<br />
Prof. <strong>Nidal</strong> <strong>Hilal</strong>, DSc<br />
Director of Centre for Clean Water Technologies, Faculty of Engineering,<br />
University of Nottingham, University Park, Nottingham NG7 2RD, UK<br />
nidal.hilal@nottingham.ac.uk<br />
Prof. Clive J. Roberts<br />
Director of Nottingham Nanotechnology <strong>and</strong> Nanoscience Centre, University<br />
of Nottingham, Nottingham NG7 2RD, UK<br />
clive.roberts@nottingham.ac.uk<br />
Dr Huabing Yin<br />
Department of Electronic <strong>and</strong> Electrical Engineering, University of Glasgow,<br />
Glasgow, UK<br />
hy@elec.gla.ac.uk<br />
Dr Vasileios Koutsos<br />
Institute for Materials <strong>and</strong> Processes, School of Engineering <strong>and</strong> Centre for<br />
Materials Science & Engineering, The University of Edinburgh, The King’s<br />
Buildings, Edinburgh EH9 3JL, UK<br />
vasileios.koutsos@ed.ac.uk<br />
Prof. P. Rhodri Williams<br />
Multidisciplinary Nanotechnology Centre, School of Engineering, Swansea<br />
University, Singleton Park, Swansea SA2 8PP, UK<br />
p.r.williams@swansea.ac.uk<br />
Dr Paul Melvyn Williams<br />
Multidisciplinary Nanotechnology Centre, School of Engineering, Swansea<br />
University, Singleton Park, Swansea SA2 8PP, UK<br />
paul.melvyn.williams@swansea.ac.uk<br />
Dr Matthew Barrow<br />
Multidisciplinary Nanotechnology Centre, School of Engineering, Swansea<br />
University, Singleton Park, Swansea SA2 8PP, UK<br />
m.s.barrow@swansea.ac.uk<br />
xv
xvi List of Contributors<br />
Dr Daniel Johnson<br />
Centre for Clean Water Technologies, Faculty of Engineering, The University<br />
of Nottingham, University Park, Nottingham, NG7 2RD, UK<br />
daniel.johnson@nottingham.ac.uk<br />
Gordon McPhee<br />
Department of Electronic <strong>and</strong> Electrical Engineering, University of<br />
Glasgow, UK<br />
gmcphee@elec.gla.ac.uk<br />
Dr Phil Dobson<br />
Department of Electronic <strong>and</strong> Electrical Engineering, University of<br />
Glasgow, UK<br />
pdobson@elec.gla.ac.uk
C H A P T E R<br />
Basic Principles of Atomic<br />
Force Microscopy<br />
Daniel Johnson, <strong>Nidal</strong> <strong>Hilal</strong> <strong>and</strong><br />
W. <strong>Richard</strong> <strong>Bowen</strong><br />
O U T L I N E<br />
. Introduction 2<br />
.2 The atomic force microscope 3<br />
.3 Cantilevers <strong>and</strong> probes 6<br />
1.3.1 Effect of Probe Geometry 7<br />
.4 Imaging modes 8<br />
1.4.1 Contact Mode Imaging 9<br />
1.4.2 Intermittent Contact (Tapping) Mode 10<br />
1.4.3 Non-Contact Mode 10<br />
1.4.4 Force Volume Imaging 11<br />
1.4.5 Force Modulation Mode 12<br />
1.4.6 Lateral/Frictional Force Mode 12<br />
.5 The afm as a force sensor 2<br />
.6 Calibration of afm microcantilevers 6<br />
1.6.1 Calibration of Normal spring Constants 16<br />
1.6.2 Calibration of Torsional <strong>and</strong> Lateral spring Constants 21<br />
.7 Colloid probes 22<br />
abbreviations <strong>and</strong> symbols 23<br />
References 24<br />
Atomic Force Microscopy in Process Engineering © 2009, Elsevier Ltd
2 1. BAsIC PRINCIPLEs OF ATOMIC FORCE MICROsCOPy<br />
. InTroduCTIon<br />
The atomic force microscope (AFM), also referred to as the scanning<br />
force microscope (SFM), is part of a larger family of instruments termed<br />
the scanning probe microscopes (SPMs). These also include the scanning<br />
tunnelling microscope (STM) <strong>and</strong> scanning near field optical microscope<br />
(SNOM), amongst others. The common factor in all SPM techniques is<br />
the use of a very sharp probe, which is scanned across a surface of interest,<br />
with the interactions between the probe <strong>and</strong> the surface being used<br />
to produce a very high resolution image of the sample, potentially to the<br />
sub-nanometre scale, depending upon the technique <strong>and</strong> sharpness of<br />
the probe tip. In the case of the AFM the probe is a stylus which interacts<br />
directly with the surface, probing the repulsive <strong>and</strong> attractive forces<br />
which exist between the probe <strong>and</strong> the sample surface to produce a highresolution<br />
three-dimensional topographic image of the surface.<br />
The AFM was first described by [1]Binnig et al. as a new technique<br />
for imaging the topography of surfaces to a high resolution. It was created<br />
as a solution to the limitations of the STM, which was able to image<br />
only conductive samples in vacuum. Since then the AFM has enjoyed<br />
an increasingly ubiquitous role in the study of surface science, as both<br />
an imaging <strong>and</strong> surface characterisation technique, <strong>and</strong> also as a means<br />
of probing interaction forces between surfaces or molecules of interest<br />
by the application of force to these systems. The AFM has a number of<br />
advantages over electron microscope techniques, primarily its versatility<br />
in being able to take measurements in air or fluid environments rather<br />
than in high vacuum, which allows the imaging of polymeric <strong>and</strong> biological<br />
samples in their native state. In addition, it is highly adaptable<br />
with probes being able to be chemically functionalised to allow quantitative<br />
measurement of interactions between many different types of<br />
materials – a technique often referred to as chemical force microscopy.<br />
At the core of an AFM instrument is a sharp probe mounted near to<br />
the end of a flexible microcantilever arm. By raster-scanning this probe<br />
across a surface of interest <strong>and</strong> simultaneously monitoring the deflection<br />
of this arm as it meets the topographic features present on the surface,<br />
a three-dimensional picture can be built up of the surface of the sample<br />
to a high resolution. Many different variations of this basic technique<br />
are currently used to image surfaces using the AFM, depending upon<br />
the properties of the sample <strong>and</strong> the information to be extracted from it.<br />
These variations include ‘static’ techniques such as contact mode, where<br />
the probe remains in constant contact with the sample, <strong>and</strong> ‘dynamic’<br />
modes, where the cantilever may be oscillated, such as with the intermittent<br />
or non-contact modes. The forces of interaction between the probe<br />
<strong>and</strong> the sample may also be measured as a function of distance by the
1.2 THE ATOMIC FORCE MICROsCOPE 3<br />
monitoring of the deflection of the cantilever, providing that the spring<br />
constant of the lever arm is sufficiently calibrated.<br />
In this chapter the basic principles of operation of an AFM will be presented,<br />
outlining the most common imaging modes <strong>and</strong> describing the<br />
acquisition of force distance measurements <strong>and</strong> techniques to calibrate<br />
cantilever spring constants.<br />
.2 The aTomIC forCe mICrosCope<br />
In Figure 1.1 the basic set-up of a typical AFM is shown. Cantilevers<br />
are commonly either V-shaped, as shown, or a rectangular, ‘diving<br />
board’ shaped. The cantilever has at its free end a sharp tip, which acts<br />
as the probe of interactions. This probe is most commonly in the form<br />
of a square-based pyramid or a cylindrical cone. A few examples of different<br />
configurations for levers <strong>and</strong> probes are shown in Figure 1.2.<br />
Commercially manufactured probes <strong>and</strong> cantilevers are predominantly<br />
of silicon nitride (the formula normally given for silicon nitride is Si 3N 4,<br />
although the precise stoichiometry may vary depending on the manufacturing<br />
process) or silicon (Si). Typically the upper surface of the cantilever,<br />
opposite to the tip, is coated with a thin reflective surface, usually of<br />
either gold (Au) or aluminium (Al).<br />
The probe is brought into <strong>and</strong> out of contact with the sample surface<br />
by the use of a piezocrystal upon which either the cantilever chip or the<br />
surface itself is mounted, depending upon the particular system being<br />
Photodetector<br />
A B<br />
Sample surface<br />
C D<br />
Light path<br />
Probe<br />
Light source<br />
y<br />
Chip<br />
Cantilever<br />
fIgure . Basic AFM set-up. A probe is mounted at the apex of a flexible Si or Si 3N 4<br />
cantilever. The cantilever itself or the sample surface is mounted on a piezocrystal which<br />
allows the position of the probe to be moved in relation to the surface. Deflection of the<br />
cantilever is monitored by the change in the path of a beam of laser light deflected from the<br />
upper side of the end of the cantilever by a photodetector. As the tip is brought into contact<br />
with the sample surface, by the movement of the piezocrystal, its deflection is monitored.<br />
This deflection can then be used to calculate interaction forces between probe <strong>and</strong> sample.<br />
z<br />
x
4 1. BAsIC PRINCIPLEs OF ATOMIC FORCE MICROsCOPy<br />
A B<br />
C D<br />
fIgure .2 Example of SEM images of different probes <strong>and</strong> cantilever types.<br />
A: pyramidal probe; B: conical high aspect ratio probe for high resolution imaging; C: two<br />
V-shaped cantilevers for contact mode imaging; D: chip with a series of beam-shaped levers<br />
of different lengths. In this case the levers are tipless to allow mounting of particles of interest<br />
for force measurements.<br />
used (these two configurations are referred to as tip-scanning or surfacescanning,<br />
respectively). Movement in this direction is conventionally<br />
referred to as the z-axis. A beam of laser light is reflected from the reverse<br />
(uppermost) side of the cantilever onto a position-sensitive photodetector.<br />
Any deflection of the cantilever will produce a change in the position of<br />
the laser spot on the photodetector, allowing changes to the deflection to<br />
be monitored. The most common configuration for the photodetector is<br />
that of a quadrant photodiode divided into four parts with a horizontal<br />
<strong>and</strong> a vertical dividing line. If each section of the detector is labelled A<br />
to D as shown in Figure 1.1, then the deflection signal is calculated by<br />
the difference in signal detected by the A � B versus C � D quadrants.<br />
Comparison of the signal strength detected by A � C versus B � D<br />
will allow detection of lateral or torsional bending of the lever. Once<br />
the probe is in contact with the surface, it can then be raster-scanned<br />
across the surface to build up relative height information of topographic<br />
features of the sample.
Force<br />
1.2 THE ATOMIC FORCE MICROsCOPE 5<br />
Intermittent<br />
contact mode<br />
Contact mode<br />
Distance<br />
Repulsion<br />
Non-contact mode<br />
Attraction<br />
fIgure .3 Diagram illustrating the force regimes under which each of the three most<br />
common AFM imaging modes operate. Contact mode operation is in the repulsive force<br />
regime, where the probe is pressed against the sample surface, causing an upwards deflection<br />
of the cantilever. Non-contact mode interrogates the long-range forces experienced<br />
prior to actual contact with the surface. With intermittent contact, or tapping mode, the<br />
probe is oscillated close to the surface where it repeatedly comes into <strong>and</strong> out of contact<br />
with the surface.<br />
This ‘optical lever’ method to detect deflection of the cantilever is the<br />
method primarily in use currently [2, 3]. However, the original design<br />
for the AFM used an STM piggy-backed onto the upper side of the AFM<br />
as a deflection sensor [1]. Whilst this allowed extremely accurate determination<br />
of the deflection, it also could detect deflection only within a<br />
very small range, which is insufficient for most purposes. In addition,<br />
other methods have been trialled in the past for this purpose including<br />
the measurement of optical interferometry effects [4] as well as fabricating<br />
levers to be able to detect deflection through a piezoresistance-based<br />
mechanism [5–14].<br />
Scanners are available in different configurations, depending upon<br />
the particular AFM employed, or the purpose for which it is required.<br />
Tube scanners consist of a hollow tube made of piezoceramic material.<br />
Depending on how electrical current is applied, the tube may extend in<br />
the z-direction or be caused to flex in either the x- or y-direction to facilitate<br />
scanning. Alternatively scanners may consist of separate piezocrystals<br />
for each movement direction. Such a configuration removes certain<br />
non-linearity problems, which may occur in the simpler tube scanners.<br />
In many commercially available AFMs, especially in older models, the<br />
movement in the x- <strong>and</strong> y-directions may be achieved by the movement<br />
of the sample rather than by the movement of the probe.
6 1. BAsIC PRINCIPLEs OF ATOMIC FORCE MICROsCOPy<br />
.3 CanTIlevers <strong>and</strong> probes<br />
Depending upon the uses required <strong>and</strong> the forces which may act upon<br />
them, cantilevers may be chosen from a large range available. Most microcantilevers<br />
used in AFMs are produced from monolithic Si 3N 4 or Si using<br />
micromachining techniques developed in the semi-conductor industry [9,<br />
15–20]. These two materials are used extensively due to the high suitability<br />
of their mechanical properties such as high yield strengths <strong>and</strong> elastic moduli<br />
[21–24]. Applications where high forces are experienced require stiff<br />
levers with probes resistant to deformation. On the contrary, where low<br />
forces are experienced or when samples are soft <strong>and</strong> easily deformed, levers<br />
with low force constants are required for both the increased force sensitivity<br />
at low forces <strong>and</strong> avoiding deforming or damaging samples. In addition to<br />
these silicon-based cantilevers <strong>and</strong> probes, the literature describes a number<br />
of other materials which have been utilised in the production of AFM<br />
levers. These include the production of diamond tips integrated into silicon<br />
levers [25, 26] or fabricated diamond levers [27], particularly useful where<br />
the probe tip may be subject to very high pressures at the apex, such as during<br />
nano-indentation measurements; quartz ‘tuning fork’ levers with polymeric<br />
tips for dynamic imaging modes [28]; metal wires, such as tungsten,<br />
as levers [4, 29, 30]; as well as more exotic materials [31, 32]. In addition,<br />
ultrasharp probes with very high aspect ratios can be manufactured by the<br />
growing of Si ‘whiskers’ on the apex of the probe to improve the resolution<br />
of samples with rough surfaces due to their relatively high aspect ratio [33,<br />
34]. In addition, a large amount of work is being undertaken to investigate<br />
the attachment of carbon nanotubes to the apex of cantilever tips to act as<br />
ultrasharp probes. Carbon nanotubes have diameters typically in the range<br />
of 1–20 nm <strong>and</strong> a very high aspect ratio making them ideal, providing that<br />
they can be attached in a robust <strong>and</strong> predictable manner [35–39].<br />
Another topic which must be considered when undertaking AFM<br />
measurements is the presence of contaminants on the probe, particularly<br />
at the tip. Commercially bought probes are placed on a sticky polymer<br />
surface, such as polydimethylsiloxane (PDMS), in plastic boxes <strong>and</strong> as a<br />
result tend to be coated in a hydrophobic contaminant layer [40]. In addition<br />
to this, once removed from packaging, the probes are likely to suffer<br />
exposure to airborne organic contaminants with the likelihood of exposure<br />
increasing over time. This can have a significant effect on imaging<br />
resolution, which depends to a great extent on the radius of curvature<br />
of the probe tip being as small as possible. Even the presence of a small<br />
amount of contamination could increase this significantly. It has also been<br />
observed that the presence of a contamination layer can increase the measured<br />
adhesion values between a probe <strong>and</strong> surface under ambient conditions<br />
[41]. The resultant increased forces between probe <strong>and</strong> sample will
also increase the effective area of contact, further decreasing the resolution<br />
of images. The presence of a contaminant layer can also adversely affect<br />
any attempt to chemically functionalise probes, preventing formation of<br />
an even layer of the desired coating material. There are various methods<br />
available to successfully clean AFM probes, which will now be described.<br />
Technologically, the simplest method is chemical cleaning by immersing<br />
the probe in an acid peroxide solution – most commonly a mixture<br />
referred to as piranha solution (usually a 7:3 ratio by volume of concentrated<br />
H 2SO 4 <strong>and</strong> 30% v/v H 2O 2, although the proportions may vary<br />
between laboratories) for a short period of time, which has been verified<br />
as effective in removing organic surface contaminants <strong>and</strong> increasing the<br />
hydrophilicity of the levers, although inorganic contaminants are not<br />
affected [42, 43]. This mixture is used in the semiconductor industry to<br />
clean photoresist <strong>and</strong> other contaminants from silicon wafers [44]. Simple<br />
cleaning by rinsing with organic solvents is insufficient to remove all the<br />
organic contaminants [43]. Piranha solution is extremely reactive with any<br />
organic material <strong>and</strong> can remove contaminants from levers <strong>and</strong> silicon<br />
surfaces in a very short space, with necessary exposure times to remove<br />
contaminants from AFM probes, which is typically less than 1 min.<br />
However, for this reason it is very corrosive if it comes into contact with<br />
any biological material including skin <strong>and</strong> must be h<strong>and</strong>led with great<br />
care by the user. In addition if it comes into contact with a relatively large<br />
quantity of organic material, such as by allowing to mix with organic solvents,<br />
the mixture may become explosive, posing a significant danger to<br />
laboratory users, with accounts of such laboratory accidents extant in the<br />
literature [45, 46]. As such, this method is not recommended if other safer<br />
cleaning methods are available <strong>and</strong> only small volumes should be produced<br />
at a time as <strong>and</strong> when required. Another commonly used method<br />
to clean AFM probes is to use ultraviolet (UV) light [47, 48]. This works<br />
by converting oxygen to form small amounts of ozone, which is then further<br />
broken down to produce highly reactive singlet oxygen, which in<br />
turn reacts with organic contaminant materials on the surfaces. For greater<br />
effectiveness, such a system is often combined with an independent ozone<br />
source to increase the amount of singlet oxygen radicals produced. Many<br />
laboratories also use plasma ashing or etching processes to remove contaminants<br />
from probes <strong>and</strong> samples [41, 42, 48–50], which has been demonstrated<br />
to be particularly effective in the removal of thin layers of organic<br />
contamination. Here a process gas, usually oxygen or argon, is ionised<br />
under a partial vacuum in a chamber containing the sample to be cleaned.<br />
.3. effect of probe geometry<br />
1.3 CANTILEvERs ANd PROBEs 7<br />
Because all measurements made using an AFM are based upon the<br />
physical interaction between the probe <strong>and</strong> the sample, it follows that the
8 1. BAsIC PRINCIPLEs OF ATOMIC FORCE MICROsCOPy<br />
shape of the probe is of fundamental importance in determining those<br />
interactions. In addition to the radius of curvature at the apex of the<br />
probe, the geometry of all of the parts of the probe which can interact<br />
with the sample are of great importance, particularly when imaging or<br />
performing indentation measurements.<br />
When imaging a sample surface, features of greatly varying geometry<br />
may be encountered. The ability to resolve these features depends<br />
upon both the sharpness of the probe tip <strong>and</strong> the aspect ratio of the<br />
probe. First, the probe sample contact area is a limit to AFM resolution.<br />
This is dependent not just upon the sharpness of the probe tip,<br />
but also upon the force with which the probe presses into the surface<br />
<strong>and</strong> the consequent mechanical deformations induced in the probe<br />
<strong>and</strong> sample. As the sample is deformed, the probe tip will become<br />
indented into the sample <strong>and</strong> a greater part of the probe surface will be in<br />
direct contact with the sample. Conversely if it presses against a hard<br />
sample with sufficient force, the probe itself may become deformed,<br />
similarly contributing to an increased interaction area. This interaction<br />
will be dependent upon the shape of the probe as well as purely on the<br />
radius of curvature found at the apex. Features present upon the surface<br />
which are smaller in size than the contact area will be unable to be<br />
successfully resolved.<br />
When asperities, which are sharper than the probe, on the surface are<br />
encountered, the image of the feature which is obtained will be based<br />
more upon the shape of the probe than upon the surface feature [51, 52]<br />
due to convolution effects. This is a problem potentially arising in all<br />
forms of SPM. This may be commonly observed when low aspect ratio<br />
probes are scanned over surfaces with high aspect ratio asperities. In<br />
effect the probe is imaged by the surface. A similar effect may be seen<br />
when the probe encounters a step edge on the sample, which is steeper<br />
than the side of the probe. In this case a broadening effect will occur<br />
where the feature under observation will appear to be wider than it actually<br />
is. These convolution effects serve to produce images which rather<br />
than being true representations of sample topography represent a composite<br />
of the topographies of both sample <strong>and</strong> probe. The finer the tip<br />
<strong>and</strong> higher the aspect ratio of the probe, will provide images which are<br />
a truer representation of the real topography of the sample.<br />
.4 ImagIng modes<br />
There are many different imaging modes available for the AFM, providing<br />
a range of different information about the sample surfaces being<br />
examined. However, for simplicity the most common modes will be<br />
considered here. In Figure 1.3 the force regimes under which the main
imaging modes occur are illustrated schematically. In this figure the<br />
interaction forces are sketched as the probe approaches <strong>and</strong> contacts the<br />
surface, with distance increasing to the right. At large separations there<br />
are no net forces acting between the probe <strong>and</strong> the sample surface. As<br />
probe <strong>and</strong> surface approach each other, attractive van der Waals interactions<br />
begin to pull the probe towards the surface. As contact is made,<br />
the net interaction becomes repulsive as electron shells in atoms in the<br />
opposing surfaces repel each other. In this figure, the repulsive forces are<br />
shown as being positive <strong>and</strong> attractive forces negative.<br />
.4. Contact mode Imaging<br />
1.4 IMAgINg MOdEs<br />
Contact mode imaging is so called because the probe remains in contact<br />
with the sample at all times. As a result, the probe–sample interaction<br />
occurs in the repulsive regime as illustrated in Figure 1.3. This is the<br />
simplest mode of AFM operation <strong>and</strong> was that originally used to scan<br />
surfaces in early instruments. There are two variations on this technique:<br />
constant force <strong>and</strong> variable force. With constant force mode, a feedback<br />
mechanism is utilised to keep the deflection, <strong>and</strong> hence force, of the<br />
cantilever constant. As the cantilever is deflected the z-height is altered<br />
to cause a return to the original deflection or ‘set point’. The change in<br />
z-position is monitored <strong>and</strong> this information as a function of the x,y-<br />
position is used to create a topographical image of the sample surface.<br />
For variable force imaging, the feedback mechanisms are switched off so<br />
that z-height remains constant <strong>and</strong> the deflection is monitored to produce<br />
a topographic image. This mode can be used only on samples which are<br />
relatively smooth with low lying surface features, but for surfaces to<br />
which it is applicable, it can provide images with a sharper resolution<br />
than constant force mode.<br />
Contact mode is often the mode of choice when imaging a hard <strong>and</strong><br />
relatively flat surface due to its simplicity of operation. However, there<br />
are several drawbacks. Lateral forces can occur when the probe traverses<br />
steep edges on the sample, which may cause damage to the probe or<br />
the sample, or also result from adhesive or frictional forces between the<br />
probe <strong>and</strong> the sample. This can also lead to a decrease in the resolution of<br />
images due to the ‘stick-slip’ movement of the probe tip over the surface.<br />
In addition, the relatively high forces with which the probe interacts with<br />
the sample can cause deformation of the sample, leading to an underestimation<br />
of the height of surface features, as well as causing an increase in<br />
the area of contact between the probe <strong>and</strong> the surface. The area of contact<br />
between the probe <strong>and</strong> the surface sets a limit to the resolution which<br />
can be achieved. Where a soft, <strong>and</strong> therefore easily deformable <strong>and</strong> easily<br />
damaged, sample is to be imaged, dynamic modes of imaging, such as<br />
intermittent contact or non-contact modes, are usually preferable.
0 1. BAsIC PRINCIPLEs OF ATOMIC FORCE MICROsCOPy<br />
.4.2 Intermittent Contact (Tapping) mode<br />
In order to overcome the limitations of contact mode imaging as mentioned<br />
earlier, the intermittent, or tapping, mode of imaging was developed<br />
[53–55]. Here the cantilever is allowed to oscillate at a value close<br />
to its resonant frequency. When the oscillations occur close to a sample<br />
surface, the probe will repeatedly engage <strong>and</strong> disengage with the surface,<br />
restricting the amplitude of oscillation. As the surface is scanned,<br />
the oscillatory amplitude of the cantilever will change as it encounters<br />
differing topography. By using a feedback mechanism to alter the<br />
z-height of the piezocrystal <strong>and</strong> maintain a constant amplitude, an image<br />
of the surface topography may be obtained in a similar manner as with<br />
contact mode imaging. In this way as the probe is scanned across the surface,<br />
lateral forces are greatly reduced compared with the contact mode.<br />
When using tapping mode in air, capillary forces due to thin layers of<br />
adsorbed water on surfaces, as well as any other adhesive forces which<br />
may be present, have to be overcome. If the restoring force of the cantilever<br />
due to its deflection is insufficient to overcome adhesion between<br />
the probe <strong>and</strong> the surface, then the probe will be dragged along the surface<br />
in an inadvertent contact mode. As a result, for this mode in air the<br />
spring constants of AFM cantilevers are by necessity several orders of<br />
magnitude greater than those used for either tapping mode in liquid or<br />
contact mode (typically in the range of 0.01–2 Nm �1 for contact mode to<br />
20–75 Nm �1 for tapping in air).<br />
As surfaces with different mechanical <strong>and</strong> adhesive properties are<br />
scanned, the frequency of oscillation will change, causing a shift in the<br />
phase signal between the drive frequency <strong>and</strong> the frequency with which<br />
the cantilever is actually oscillating [56, 57]. This phenomenon has been<br />
used to produce phase images alongside topographic images, which are<br />
able to show changes in the material properties of the surfaces being<br />
investigated. However, whilst the qualitative data provided by the phase<br />
images are useful, it is difficult to extract quantitative information from<br />
them because they are a complex result of a number of parameters including<br />
adhesion, scan speed, load force, topography <strong>and</strong> the material, especially<br />
elastic, properties of the sample <strong>and</strong> probe [57, 58].<br />
.4.3 non-Contact mode<br />
In non-contact mode imaging, the cantilever is again oscillated as in<br />
intermittent contact mode, but at much smaller amplitude. As the probe<br />
approaches the sample surface, long-range interactions, such as van der<br />
Waals <strong>and</strong> electrostatic forces, occur between atoms in the probe <strong>and</strong> the<br />
sample. This causes a detectable shift in the frequency of the cantilever’s
oscillations. Detection of the shift in phase between the driving <strong>and</strong> oscillating<br />
frequencies allows the z-positioning of the cantilever to be adjusted<br />
to allow the cantilever to remain out of contact with the surface by the<br />
operation of a feedback loop [59]. Because the probe does not contact the<br />
surface in the repulsive regime, the area of interaction between the tip<br />
<strong>and</strong> the surface is minimised allowing potentially for greater surface resolution.<br />
As a result in this mode, it is imaging which is best able to achieve<br />
true atomic resolution, when examining a suitable surface under suitable<br />
conditions. However, in practice obtaining images of a high quality is a<br />
more daunting prospect than intermittent contact mode. When imaging<br />
in air all but the most hydrophobic regions of surfaces will have a significant<br />
water layer, which may be thicker than the range of the van der<br />
Waals forces being probed. This combined with the low oscillation amplitude<br />
will mean that the probe will be unable to detach from the water<br />
layer easily, degrading imaging resolution.<br />
.4.4 force volume Imaging<br />
1.4 IMAgINg MOdEs<br />
The force volume imaging mode is a combination of conventional<br />
imaging with the measurement of force distance curves (see Section 2.2<br />
for explanation of these measurements). For each pixel of the image, a<br />
force distance curve is obtained by bringing the probe into <strong>and</strong> then out<br />
of contact with the surface <strong>and</strong> simultaneously recording the deflection<br />
of the lever as a function of the z-directional translation, so that concurrently<br />
with obtaining a conventional topographic (height) image, information<br />
about how interaction forces between the probe <strong>and</strong> the sample<br />
vary with the sample topography is obtained. By plotting an image<br />
scaled to show the lowest force value for each pixel, an adhesion map of<br />
the surface can be obtained [60]. This is useful for highlighting different<br />
adhesive properties of different surface domains [60–62], which can be<br />
particularly useful if the probe itself carries a tailored chemical functionality<br />
[63, 64]. In addition plots of the relative elasticity of different surface<br />
domains can be produced, if working with sufficiently soft samples<br />
such as polymer layers [65] or surface immobilised cells [63]. Where force<br />
measurements are to be obtained between the probe <strong>and</strong> a relatively<br />
rough surface, a high variability between curves may be observed due<br />
to surface asperities, causing a great variation in the contact area from<br />
one point to another. If this is the case then the force volume mode is a<br />
useful way to obtain the large numbers of force curves needed over an<br />
area. As obtaining a large number of force curves at the same time is very<br />
memory intensive, force volume images are typically lower in resolution,<br />
in terms of the number of pixels in an image than the images obtained in<br />
other AFM modes.
2 1. BAsIC PRINCIPLEs OF ATOMIC FORCE MICROsCOPy<br />
.4.5 force modulation mode<br />
Force modulation mode combines some of the aspects of both contact<br />
<strong>and</strong> dynamic modes of imaging. During the operation of the force<br />
modulation mode of AFM, either the cantilever or the sample is oscillated<br />
sinusoidally in the z-direction, while raster-scanning the probe<br />
across the sample surface whilst in constant contact [66–70]. The amplitude<br />
<strong>and</strong> phase of these oscillations is monitored simultaneously with<br />
the creation of a topographic image of the surface. This allows changes in<br />
the mechanical compliance of the surface to be monitored <strong>and</strong> compared<br />
with changes in the topography. From the knowledge of the stiffness of<br />
the cantilever, the geometry of the probe apex <strong>and</strong> its area of contact with<br />
the sample <strong>and</strong> the utilisation of an appropriate contact mechanics model<br />
such as the Hertz or the Johnson, Kendall <strong>and</strong> Roberts (JKR) models, it<br />
is possible to extract useful quantitative information about the material<br />
properties of the surface, such as the elastic modulus. The ability of this<br />
technique to monitor differences in the surface mechanical properties<br />
between different domains of surfaces to the high resolution obtainable<br />
with the AFM is of much use in the characterisation of materials engineered<br />
on the nano-scale.<br />
.4.6 lateral/frictional force mode<br />
In lateral force mode, the forces exerted upon the probe tip in the lateral<br />
(x) direction as it is scanned across a surface are recorded simultaneously<br />
with topography. This is of particular interest for obtaining<br />
quantitative measurements of the frictional forces felt between the probe<br />
<strong>and</strong> the sample [29, 30, 71–74] <strong>and</strong> of much interest in the field of nanotribology,<br />
where the frictional <strong>and</strong> wearing properties of materials used to<br />
construct micro-machines is of great importance due to their high surface<br />
area to volume ratio. To extract quantitative data from the measurements,<br />
a number of variables need to be accounted for, such as the normal force<br />
applied by the probe tip, the lateral spring constant of the cantilever (see<br />
Section 1.4.2), the geometry of the tip apex (particularly its area of contact<br />
with the surface) <strong>and</strong> the sensitivity of the optical lever to the torsional<br />
bending of the lever arm.<br />
.5 The afm as a forCe sensor<br />
One of the major applications of the AFM is in the quantitative measurement<br />
of interaction forces between either the probe tip, or an attached<br />
particle replacing the tip, <strong>and</strong> a sample surface. This technique has been<br />
employed to examine a wide variety of systems including the mechanical
1.5 THE AFM As A FORCE sENsOR 3<br />
properties of materials on the micro- <strong>and</strong> nano-scales [75–80] of interest<br />
for characterising nano-engineered materials; adhesion between surfaces<br />
[80–85]; attractive <strong>and</strong> repulsive surface forces, such as van der Waals <strong>and</strong><br />
electrostatic double layer forces [86–89] both of interest when studying<br />
the properties of colloidal particles; <strong>and</strong> to probe the mechanical properties<br />
<strong>and</strong> kinetics of bond strength of biomolecules [90–93].<br />
As the tip of the cantilever is brought into <strong>and</strong> out of contact with a<br />
surface, a force curve is generated, describing the cantilever deflection<br />
(or force) as a function of distance. A typical force curve is illustrated in<br />
Figure 1.4. Raw data are plotted as displacement of the z-piezo on the<br />
abscissa, whilst cantilever deflection is plotted as the signal on the photodetector<br />
(commonly either as voltage V or sometimes as current A)<br />
on the ordinate. As the cantilever begins its approach (described by the<br />
red trace), it is away from the surface <strong>and</strong> hence there is no detection of<br />
change in force (point 1 in the figure) – the cantilever is said to be at its<br />
‘free level’, i.e. at this point there are no net forces acting on it (assuming<br />
that the probe is not travelling fast enough for hydrodynamic drag<br />
forces to have a significant effect). As the probe comes into close proximity<br />
with the cantilever, long-range forces may cause interaction between<br />
the probe <strong>and</strong> the objective surface. Repulsive forces will cause the lever<br />
to deflect upwards <strong>and</strong> away from the surface, whereas attractive forces<br />
will deflect the lever downwards, towards the surface. If the gradient of<br />
attractive forces is less than the stiffness of the lever, then the probe will<br />
momentarily be deflected downwards, before re-equilibrating at its free<br />
level due to the restoring force stored in the lever. If the probe reaches<br />
a point where the gradient of attractive forces exceeds the stiffness of<br />
the cantilever, then the cantilever will be rapidly deflected downwards<br />
allowing the probe to touch the surface in a ‘snap-in’ or ‘jump-to-<br />
contact’. In the absence of attractive surface forces, this jump-to-contact<br />
will not be seen. When the cantilever makes hard contact with the surface,<br />
it is deflected upwards due to repulsion between electron shells of<br />
atoms in the opposing material surfaces, <strong>and</strong> a positive force is observed<br />
(point 2). The cantilever is then retracted <strong>and</strong> initially follows the path<br />
of the approach trace in the contact region. The cantilever often remains<br />
attached to the surface by adhesive forces which results in a downwards<br />
deflection of the cantilever as the probe retracts away from the surface,<br />
causing a hysteresis between the trace <strong>and</strong> the retrace. Eventually the<br />
separation force becomes sufficient to overcome the adhesion between<br />
the probe tip <strong>and</strong> the surface, <strong>and</strong> the cantilever snaps back to its initial<br />
free level position (point 3).<br />
This behaviour results in a curve of deflection (measured as raw signal)<br />
versus displacement of the piezo in the z-direction. When the surface<br />
being pressed against is hard <strong>and</strong> does not undergo significant deformation,<br />
the z-movement will be equal to the deflection of the cantilever. As a
4 1. BAsIC PRINCIPLEs OF ATOMIC FORCE MICROsCOPy<br />
Deflection (nA)<br />
A<br />
Force (nN)<br />
B<br />
60<br />
50<br />
40<br />
30<br />
20<br />
10<br />
0<br />
0 50 100 150 200 250 300 350 400<br />
–10<br />
–20<br />
–30<br />
35<br />
30<br />
25<br />
20<br />
15<br />
10<br />
5<br />
0<br />
–100 –50 0 50 100 150 200 250 300<br />
–5<br />
–10<br />
2.<br />
z-piezo translation (nm)<br />
Distance (nm)<br />
fIgure .4 An example of force curve. Raw data are shown in (a) plotted as displacement<br />
of the z-piezo versus deflection as measured on the photosensitive detector.<br />
Calculation of the cantilever stiffness <strong>and</strong> sensitivity of the optical lever set-up allow the<br />
raw data to be converted into probe–sample separation distance versus the force, with converted<br />
force curve shown in (b). By convention, for AFM force curves, attractive forces are<br />
portrayed as negative <strong>and</strong> repulsive forces as positive.<br />
3.<br />
1.
1.5 THE AFM As A FORCE sENsOR 5<br />
consequence, the slope obtained from contact with an unyielding surface<br />
provides the sensitivity of the optical lever system. By dividing the raw<br />
deflection data by this sensitivity value, it can be converted into an actual<br />
deflection distance. This sensitivity value is essential for the calculation<br />
of force values from raw deflection data. This deflection value (distance)<br />
can also then be subtracted from the z-piezo displacement to give the<br />
actual distance travelled by the probe. An alternative method of finding<br />
the optical lever sensitivity without needing to make hard contact with<br />
the surface was suggested by Higgins et al. [94] (Section 1.4).<br />
Within operational limits the AFM cantilever behaves as a linear, or<br />
‘Hookean’, spring. As a result the magnitude of the deflection of the cantilever<br />
can be used to calculate the force which is exerted on the cantilever<br />
using Hooke’s law:<br />
F � � kx<br />
(1.1)<br />
where F is force (N), x the deflection of the cantilever (m) <strong>and</strong> k the spring<br />
(or force) constant of the cantilever (N m �1 ), which essentially represents<br />
the stiffness of the cantilever. This spring constant is dependent upon the<br />
physical properties of the lever. This is apparent from the following relation,<br />
used to describe a rectangular, ‘diving board’ shaped lever as shown<br />
in 1. 5 [16, 95]:<br />
3<br />
Et w<br />
k � 3<br />
4l<br />
(1.2)<br />
where E is the Young’s modulus of the lever <strong>and</strong> t, w <strong>and</strong> l the thickness,<br />
width <strong>and</strong> length of the lever, respectively. However, it must be borne in<br />
mind that none of the diving board levers are perfectly rectangular, due<br />
to shaping of the ends of the beams <strong>and</strong> imperfections in the manufacturing<br />
process. As a result this equation will give only an approximate value<br />
for the stiffness of a rectangular cantilever. In the next section the need<br />
for the methods used to effectively measure the stiffness of a cantilever<br />
to be used for force measurements, whether it is rectangular or V-shaped,<br />
<strong>and</strong> the reasons for variability in k between cantilevers are addressed in<br />
greater detail.<br />
For different applications, cantilevers with different spring constants<br />
may be needed. For instance, for intermittent contact mode in air, particularly<br />
stiff levers are needed to overcome capillary forces, whereas for<br />
measurement of weak interaction forces, very soft levers are needed for<br />
their increased force sensitivity. The most convenient ways of producing<br />
this variation are by altering the length <strong>and</strong> or thickness of levers during<br />
production in order to increase or decrease the lever stiffness.
6 1. BAsIC PRINCIPLEs OF ATOMIC FORCE MICROsCOPy<br />
.6 CalIbraTIon of afm mICroCanTIlevers<br />
.6. Calibration of normal spring Constants<br />
For the accurate measurement of forces, the spring constant of the cantilever<br />
needs to be known. In particular, for forces normal to the surfaces<br />
of interest, this is the spring constant which governs the relationship<br />
between force <strong>and</strong> deflection in the z-direction, as opposed to the lateral<br />
<strong>and</strong> torsional spring constants (see Section 1.4.2). Although cantilevers<br />
are supplied with a manufacturers’ ‘nominal’ value, the actual value can<br />
vary to a high degree, mostly due to variations in the thickness of the<br />
levers <strong>and</strong> defects in the material of the cantilevers themselves. The cubic<br />
relationship between thickness <strong>and</strong> the spring constant seen in equation<br />
(1.2) means that small variations in thickness can cause significant variations<br />
in k. Because of this variability, for force experiments the cantilevers<br />
to be used need to be calibrated to determine a more accurate value of k.<br />
There are now a large number of methods by which the spring constant<br />
can be calculated, each with their own advantages <strong>and</strong> disadvantages.<br />
Four of the most extensively used approaches are described later.<br />
A number of methods exist which involve calculating spring constants<br />
based upon the dimensions <strong>and</strong> geometry of the lever. Whilst calculations<br />
for rectangular cantilevers, such as in equation (1.2), are relatively<br />
straightforward, for V-shaped cantilevers, approximations are most often<br />
used based upon a simplification of their geometry, e.g. Sader approximated<br />
a V-shaped cantilever to two parallel rectangular beams [96]. This<br />
resulted in the following equation to describe a V-shaped lever:<br />
3<br />
Et w ⎧ 3<br />
4w<br />
⎫<br />
k � cos � 1 � 3 � 2<br />
3 ⎨<br />
⎪ ( cos � )<br />
3<br />
⎬<br />
⎪<br />
2l<br />
⎩<br />
⎪ b<br />
⎭<br />
⎪<br />
�1<br />
(1.3)<br />
where � is the inside angle between the two arms of the V-shaped lever,<br />
w the width of each of the lever arms parallel to the base of the lever <strong>and</strong><br />
b the outer width of the base of the lever (see Figure 1.5 for diagrammatic<br />
explanation of the dimensions). This equation, as well as equation<br />
(1.2), assumes that the point of loading of force on the cantilever will be<br />
at the very apex. As the probe tip itself, where the loading of force actually<br />
occurs, is often sited a short distance from the very end, this needs to<br />
be taken into account. A simple correction may be applied based on the<br />
length of the lever <strong>and</strong> the distance of the probe from the end of the lever<br />
[96–98]:<br />
k � k ( l/ ∆l) 3 (1.4)<br />
c m<br />
�
l<br />
1.6 CALIBRATION OF AFM MICROCANTILEvERs 7<br />
where k m is the uncorrected calculated spring constant value, k c the corrected<br />
value <strong>and</strong> �l the distance of the centre of the base of the probe<br />
from the apex of the lever.<br />
The problem with calculating spring constants purely from the measured<br />
dimensions of the lever <strong>and</strong> nominal values for the Young’s modulus is primarily<br />
that during cantilever manufacture variability in the material properties,<br />
particularly the Young’s modulus, of cantilevers can occur largely due<br />
to variations in the morphology of the silicon nitride [23]. In addition, accurate<br />
determination of lever thickness is not always very practicable. Making<br />
accurate measurements for every lever used in experiments by SEM is time<br />
consuming <strong>and</strong> not necessarily convenient. By measuring the resonance<br />
behaviour of cantilevers, variability in the material properties of the cantilevers<br />
can be at least to some extent taken into account. This leads to a more<br />
reliable calculation for the spring constant of a cantilever surrounded by a<br />
fluid environment, such as air, from the following relationship [99, 100]:<br />
2 2<br />
f i f f<br />
k � 0. 1906�<br />
b lQΓ<br />
( � ) �<br />
(1.5)<br />
where � f is the density of the surrounding fluid; Q the quality factor (a<br />
measure of the sharpness of the resonance peak); � i the imaginary component<br />
of the hydrodynamic function, dependent upon the Reynolds<br />
number of the fluid; <strong>and</strong> � f the fundamental resonance frequency of the<br />
cantilever. Although this approach is reliable for calibrating rectangular<br />
levers, there are not currently any reliable approximations to allow this<br />
method to be used for the commonly used V-shaped cantilevers.<br />
Another method, developed by Clevel<strong>and</strong> et al. [95], requires the<br />
attachment of known masses, such as tungsten spheres, to the end of the<br />
cantilever whilst monitoring the resultant resonant frequency change.<br />
Measuring the position of the fundamental resonance peak of the<br />
cantilever before <strong>and</strong> after the addition of the sphere allows the following<br />
relationship to be used to calculate the spring constant:<br />
2 M<br />
k � ( 2π<br />
) (1.6)<br />
( 1/ v ) � ( 1/<br />
v )<br />
1 2<br />
w<br />
0 2<br />
t<br />
fIgure .5 Diagram<br />
showing relevant dimensions<br />
on a beam-shaped<br />
AFM cantilever.
8 1. BAsIC PRINCIPLEs OF ATOMIC FORCE MICROsCOPy<br />
where M is the added mass, k the cantilever spring constant <strong>and</strong> v 0 <strong>and</strong> v 1<br />
the unloaded <strong>and</strong> loaded resonant frequencies.<br />
Although this method can produce a value for the cantilever<br />
spring constant to a high degree of accuracy, there are some problems.<br />
Attaching the bead to the cantilever, especially if attached using glue,<br />
can be a destructive process, rendering the lever unusable for further<br />
experiments. This means that calibration must be carried out at the end<br />
of experimental measurements. However, with sufficient care spheres<br />
may be attached in air using capillary adhesion forces alone. In addition<br />
errors may occur from the incorrect placement of the sphere or uncertainties<br />
in the mass of the sphere. If the sphere is placed a short distance<br />
away from the end of the lever, then the value obtained from this<br />
method will be high, as k is inversely proportional to the cube of the<br />
cantilever length l. Again, this can be simply accounted for by utilizing<br />
equation (1.4).<br />
The masses of spheres ordinarily used in this technique are on the<br />
order of a nanogram, making accurate weighing problematic. As such,<br />
masses are generally estimated from the size <strong>and</strong> density of the spheres,<br />
which may in turn lead to measurement errors. To allow for this, measurements<br />
may be made using several spheres of different masses. A plot of M<br />
versus (2�v 1) �2 can then be made, which will have a slope equal to the<br />
spring constant of the cantilever [95, 101]. The advantages of this method<br />
are that the measurements are independent of the cantilever geometry<br />
<strong>and</strong> material properties, <strong>and</strong> many commercially available AFMs have the<br />
capability to measure cantilever resonant frequencies.<br />
Another technique commonly used to quantify cantilever spring<br />
constants is the so-called ‘thermal method’ devised by Hutter <strong>and</strong><br />
Bechhoefer [102, 103]. Here the area of the fundamental resonant peak<br />
of the cantilever under ambient thermal excitation, when not in the presence<br />
of a surface, can be used to directly calculate the spring constant of<br />
the cantilever. The mean square deflection of the cantilever, x 2 , due to<br />
thermal fluctuations can be related to the spring constant thus, assuming<br />
an idealised spring behaviour:<br />
kBT k �<br />
2<br />
x<br />
(1.7)<br />
where k B <strong>and</strong> T are Boltzmann’s constant (1.38 � 10 �23 J K �1 ) <strong>and</strong> absolute<br />
temperature, respectively, together representing the thermal energy<br />
of the system. Once other noise sources are subtracted from the background,<br />
the area of the fundamental resonance peak will be equal to the<br />
mean square displacement. However, as the cantilever is not an ideal
spring, equation (1.7) is insufficient to describe its behaviour, <strong>and</strong> a number<br />
of other factors may need to be taken into account [103–107]:<br />
2<br />
0. 8174s kBT 1 � 3Dtan / 2l<br />
k �<br />
P 1 � 2Dtan<br />
/ l<br />
cos<br />
⎛ ϕ ⎞<br />
⎜<br />
ϕ<br />
⎝⎜<br />
ϕ ⎠⎟<br />
(1.8)<br />
where D is the tip height, s the sensitivity (in V or A m �1 ), <strong>and</strong> P the positional<br />
noise power of the fundamental resonance peak (the area under<br />
the fundamental resonance peak in the power spectral density (PSD)<br />
curve), obtained from a plot of the thermal power spectrum. Thus, with<br />
a spectrum analyser <strong>and</strong> appropriate software available with a number<br />
of commercially available AFM instruments, spring constant calibration is<br />
relatively straightforward as well as being relatively non-destructive.<br />
Higgins et al. [94] suggested a novel way of finding the optical lever<br />
sensitivity by combining this thermal method with those of Sader. By<br />
using Sader’s method to determine a spring constant for the cantilever,<br />
the method of Hutter <strong>and</strong> Bechhoefer could then be used to back-<br />
calculate the optical lever sensitivity without the need to make a hard contact<br />
with a stiff surface. This method is potentially of use where a hard<br />
contact is undesirable, for instance when the probe is chemically functionalised<br />
or when a particle made of some deformable material, which is likely<br />
to significantly deform under measurement stresses, is used as a probe.<br />
A very simple <strong>and</strong> straightforward method of calibrating the cantilever<br />
spring constant is to press the cantilever to be calibrated against another,<br />
reference, cantilever of known k (Figure 1.6). This could be either a macroscopic<br />
lever [108] or another AFM microcantilever [109]. Reference<br />
cantilevers can be obtained either commercially or by using another calibration<br />
method or combination of other methods to determine k to a high<br />
precision. When the two levers are pressed together, the slope of the contact<br />
region on the force curve will be the result of the deflection of both<br />
levers. As a result, comparison of this slope with the slope obtained when<br />
b<br />
w<br />
1.6 CALIBRATION OF AFM MICROCANTILEvERs<br />
w ′<br />
t<br />
α<br />
l ′<br />
l<br />
2<br />
fIgure .6 Diagrammatic<br />
representation of a<br />
V-shaped AFM cantilever.
20 1. BAsIC PRINCIPLEs OF ATOMIC FORCE MICROsCOPy<br />
pressed against a hard surface, which will not appreciably deform under<br />
the pressure exerted upon it, will allow the calculation of the unknown<br />
spring constant, providing that the same optical lever set-up <strong>and</strong> hence<br />
its sensitivity remains unchanged between each set of measurements:<br />
k � k<br />
c ref<br />
⎛ �hard � � ⎞<br />
⎜<br />
ref<br />
⎜<br />
⎝⎜<br />
�ref � cosϕ<br />
⎠⎟<br />
⎛<br />
�<br />
⎞<br />
⎜ hard<br />
kc � k ⎜ ref ⎜ � 1<br />
⎝⎜<br />
�ref<br />
⎠⎟<br />
(1.9)<br />
(1.10)<br />
where k c <strong>and</strong> k ref are the spring constants of the unknown <strong>and</strong> known<br />
reference levers, � hard <strong>and</strong> � ref the gradients of the contact regions of force<br />
curves against a hard surface <strong>and</strong> the reference lever, respectively, ϕ the<br />
angle between the two levers.<br />
Typically levers are mounted onto the AFM with an in-built tilt angle of<br />
approximately 10–12°, which is liable to give a value of k varying from the<br />
true value by less than the experimental uncertainty. In addition, if care is<br />
taken to maintain the same angle between the lever <strong>and</strong> the experimental<br />
sample when calibrating the lever, then the apparent spring constant calculated<br />
by this method will be identical to the effective spring constant. As such<br />
the term cos ϕ can be ignored (equation (1.10)). This is a quick <strong>and</strong> simple<br />
method to use <strong>and</strong> can be carried out where the instrumentation being used<br />
does not allow the measurement of the resonance spectrum of the cantilever.<br />
One area where caution must be taken with the reference lever method<br />
is in the precise positioning of the two levers (Figure 1.7). As seen from<br />
equation (1.2), k will vary inversely in relation to l 3 . This means that if the<br />
cantilever to be calibrated overlaps with the reference lever, effectively<br />
reducing the length of the reference, the measurements obtained will be<br />
as though taken against a stiffer reference lever, leading to an underestimation<br />
of k c. Another important factor to consider is that the stiffness of<br />
the unknown cantilever <strong>and</strong> the reference must be similar in order to get<br />
a truly accurate result. If one cantilever is much stiffer than the other, then<br />
the slope obtained from a force curve of the cantilevers pressed together<br />
k ref<br />
k c<br />
ϕ<br />
fIgure .7 Static def-<br />
lection of a cantilever of<br />
unknown spring constant<br />
against a reference cantilever.<br />
The slope of the contact<br />
region of the resultant<br />
force curve is dependent<br />
upon the stiffness of the two<br />
levers combined as well as<br />
the angle between them.
will be dominated by the deflection of the softer lever. It has been suggested<br />
that one lever should not have a spring constant greater than the<br />
other by more than a factor of three [109] for this method to be effective.<br />
This is only a selection of some of the more commonly used techniques.<br />
There are a large number of other methods used to determine<br />
spring constants of AFM cantilevers listed in the literature. These include,<br />
in no especial order, the measurement of the dynamic response of cantilevers<br />
with colloidal spheres attached in a viscous fluid [110, 111]; calculating<br />
k of levers on a chip, based on geometry, compared with a lever on<br />
the same chip calibrated by another method [97]; an alternative ‘thermal<br />
method’ with k calculated from the resonant frequency, Q factor <strong>and</strong> the<br />
squared resonance amplitude [112]; <strong>and</strong> measuring the static deflection<br />
of an AFM cantilever due to a known end-loaded mass [113].<br />
.6.2 Calibration of Torsional <strong>and</strong> lateral spring Constants<br />
To extract quantitative data from measurements of lateral forces experienced<br />
by the probe when scanning across a surface, such as in friction<br />
force microscopy, then the stiffness of the cantilever in the lateral or torsional<br />
mode needs to be ascertained. However, this is much less straightforward<br />
than calibrating the normal spring constant of the lever. For<br />
frictional measurements, as well as knowledge of the normal <strong>and</strong> torsional<br />
spring constants, knowledge of the lateral response of the deflection<br />
sensor <strong>and</strong> the geometry, height <strong>and</strong> material properties of the probe<br />
at the region of contact with the sample need to be ascertained.<br />
At this point the difference between the torsional <strong>and</strong> lateral stiffnesses<br />
of the cantilevers must be made clear. The torsional spring constant k � is the<br />
resistance to rotation along the major axis of the cantilever. The lateral spring<br />
constant k lat on the contrary is the resistance of the lever to forces experienced<br />
laterally at the apex of the probe tip, producing a rotation at the base<br />
of the probe. The two are related by the following simple formula [114]:<br />
kΦ<br />
klat<br />
� 2<br />
h<br />
(1.11)<br />
where h is the height of the probe (usually in the region of 3 μm for most<br />
imaging probes).<br />
Sader described equations to calculate the approximate k� for both<br />
beam-shaped <strong>and</strong> V-shaped cantilevers [115] from their geometries.<br />
Torsional stiffness for a beam-shaped cantilever is:<br />
k<br />
φ<br />
1.6 CALIBRATION OF AFM MICROCANTILEvERs 2<br />
⎧⎪<br />
⎛l<br />
� l<br />
3<br />
tanh ⎜<br />
∆ ⎞ ⎫<br />
⎪<br />
⎪<br />
� v<br />
Et w ⎪<br />
6( 1 ) ⎪<br />
⎝⎜<br />
w<br />
⎠⎟<br />
w ⎪<br />
� ⎨<br />
⎪1<br />
�<br />
⎬<br />
⎪<br />
6( 1 + v)( l − ∆l)<br />
⎪<br />
6( 1 � v)<br />
l � ∆l⎪<br />
⎪<br />
⎪<br />
⎩⎪<br />
⎭⎪<br />
�1<br />
(1.12)
22 1. BAsIC PRINCIPLEs OF ATOMIC FORCE MICROsCOPy<br />
<strong>and</strong> for torsional forces experienced by a V-shaped cantilever:<br />
k<br />
φ �<br />
3<br />
2<br />
Et w w ⎛2w<br />
l ⎞ 3⎛<br />
b⎞<br />
1<br />
1 � 2log⎜<br />
� 2 � ⎜<br />
3( 1 � v) l b ⎝⎜<br />
b ∆l⎠⎟<br />
8 ⎝⎜<br />
l ⎠⎟<br />
1 �� b<br />
⎧⎪<br />
⎡<br />
2<br />
⎛ ⎞ ⎤⎫<br />
⎪<br />
⎪<br />
⎪ ⎢<br />
⎜<br />
⎥⎪<br />
⎪ ⎢<br />
⎜<br />
⎥<br />
⎪<br />
⎨<br />
⎪ ⎢<br />
⎜<br />
⎢<br />
⎜<br />
⎥<br />
⎜ 2<br />
⎢<br />
⎜<br />
⎥<br />
⎬<br />
⎪<br />
⎪<br />
⎪<br />
⎪<br />
2<br />
⎢<br />
⎝⎜<br />
4l<br />
⎠⎟<br />
⎥⎪<br />
⎪<br />
⎥<br />
⎪<br />
⎩⎪<br />
⎣<br />
⎦⎭⎪<br />
−1<br />
(1.13)<br />
where �l is the distance of the base of the probe tip from the apex of the<br />
lever.<br />
For relevant dimensions of the levers see Figures 1.5 <strong>and</strong> 1.6. In terms<br />
of measured interactions, k lat can be defined in similar terms to Hooke’s<br />
law for the normal spring constant of the lever [73]:<br />
Flat � klat∆ x<br />
(1.14)<br />
where F lat is the force experienced by the tip <strong>and</strong> �x the lateral movement<br />
of the tip along the x-direction (i.e. at right angles to the major axis<br />
of the cantilever).<br />
.7 ColloId probes<br />
By attaching a microsphere to the end of a tipless cantilever, the geometry<br />
of interactions between the probe <strong>and</strong> the surface can be greatly simplified,<br />
allowing the AFM to be used to probe surface forces, much akin to the<br />
surface force apparatus (SFA). Microparticles can also be used as probes,<br />
which will allow particle to particle adhesion forces to be measured.<br />
An example of a colloid probe is illustrated in Figure 1.8. Here a scanning<br />
fIgure .8 SEM image<br />
of a colloid probe created<br />
by the attachment of a silicon<br />
dioxide sphere to the<br />
apex of an AFM microcantilever<br />
using an epoxy<br />
resin. The scale bar shown<br />
is 1 �m long, with the particle<br />
approximately 5 �m in<br />
diameter.
ABBREvIATIONs ANd syMBOLs 23<br />
electron microscope (SEM) image shows a 5-μm diameter silica bead<br />
attached with glue close to the apex of a st<strong>and</strong>ard AFM microcantilever.<br />
The first reported use of an AFM with a colloidal probe was by Ducker<br />
et al. [116, 117] who attached a 3.5-μm silica sphere to an AFM cantilever<br />
<strong>and</strong> used it to measure forces between the sphere <strong>and</strong> a silica surface as<br />
a function of electrolyte concentration <strong>and</strong> pH. Since then this technique<br />
has been used to probe the interaction forces between various materials<br />
<strong>and</strong> surfaces including silicates <strong>and</strong> other inorganic materials [80, 83, 117,<br />
118], protein- <strong>and</strong> polymer-coated beads <strong>and</strong> surfaces [119–122], membrane-fouling<br />
materials <strong>and</strong> membranes [123], biological cells <strong>and</strong> surfaces<br />
[124, 125]; between drug particles which are important in powder<br />
formulations [81, 84, 87]; <strong>and</strong> for probing the rheological properties of<br />
liquids [110, 126]. See Chapter 2 for more detail on the preparation of colloid<br />
probes.<br />
abbrevIaTIons <strong>and</strong> symbols<br />
AFM Atomic force microscopy/microscope<br />
b Outer width of the base of V-shaped cantilever m<br />
E Young’s modulus N m�2 F Force normal to sample surface N<br />
Flat Lateral forces N<br />
JKR Johnson, Kendall <strong>and</strong> Roberts theory<br />
k Spring constant of cantilever N m�1 kB Boltzmann’s constant (1.38 � 10�23 ) J K�1 kc Corrected k N m�1 klat Lateral spring constant N m�1 km Uncorrected k N m�1 kref Reference cantilever spring constant N m�1 k � Torsional spring constant N m �1<br />
l Cantilever length m<br />
M Mass added to cantilever kg<br />
P Positional noise power of fundamental resonant peak<br />
PDMS Polydimethylsiloxane<br />
Q Quality factor of cantilever –<br />
SPM Scanning probe microscopy<br />
STM Scanning tunnelling microscopy/microscope<br />
t Cantilever thickness m
24 1. BAsIC PRINCIPLEs OF ATOMIC FORCE MICROsCOPy<br />
T Absolute temperature K<br />
v 0 Unloaded resonant frequency Hz<br />
v 1 Loaded resonant frequency Hz<br />
w Cantilever width m<br />
x Deflection of cantilever m<br />
� Inside angle of V-shaped cantilever °<br />
� i Imaginary component of hydrodynamic function –<br />
� hard Contact slope versus hard surface nm V �1<br />
� ref Contact slope measured versus reference cantilever nm V �1<br />
� f Density of surrounding fluid Pa s<br />
ϕ Angle between cantilever <strong>and</strong> reference lever °<br />
� f Fundamental resonant frequency of lever Hz<br />
references<br />
[1] G. Binnig, C.F. Quate, C. Gerber, Atomic force microscope, Phys. Rev. Lett. 56 (9)<br />
(1986) 930–933.<br />
[2] S. Alex<strong>and</strong>er, L. Hellemans, O. Marti, J. Schneir, V. Elings, P.K. Hansma, M. Longmire,<br />
J. Gurley, An atomic resolution atomic force microscope implemented using an optical<br />
lever, J. Appl. Phys. 65 (1) (1988) 164–167.<br />
[3] G. Meyer, N.B. Amer, Novel optical approach to atomic force microscopy, Appl. Phys.<br />
Lett. 53 (12) (1988) 1045–1047.<br />
[4] Y. Martin, C.C. Williams, H.K. Wickramasinghe, Atomic force microscope – force mapping<br />
<strong>and</strong> profiling on a sub 100-Å scale, J. Appl. Phys. 61 (10) (1987) 4723–4729.<br />
[5] R. Jumpertz, A.v.d. Hart, O. Ohlsson, F. Saurenbach, J. Schelten, Piezoresistive sensors<br />
on AFM cantilevers with atomic resolution, Microelectronic Engineering, 41/42 (1998)<br />
441–444.<br />
[6] P.A. Rasmussen, J. Thaysen, S. Bouwstra, A. Boisen, Modular design of AFM probe<br />
with sputtered silicon tip, Sensors Actuat. A 92 (2001) 96–101.<br />
[7] M. Tortonese, R.C. Barrett, C.F. Quate, Atomic resolution with an atomic force microscope<br />
using piezoresistive detection, Appl. Phys. Lett. 62 (8) (1993) 834–836.<br />
[8] T. Gotszalk, P. Grabiec, F. Shi, P. Dumania, P. Hudek, I.W. Rangelow, Fabrication of<br />
multipurpose AFM/SCM/SEP microprobe with integrated piezoresistive deflection<br />
sensor <strong>and</strong> isolated conductive tip, Microelectron. Eng. 41/42 (1998) 477–480.<br />
[9] P.-F. Indermuhle, G. Schurmann, G.-A. Racine, N.F. De Rooij, Fabrication <strong>and</strong> characterization<br />
of cantilevers with integrated sharp tips <strong>and</strong> piezoelectric elements for<br />
actuation <strong>and</strong> detection for parallel applications, Sensors Actuat. A 60 (1997) 186–190.<br />
[10] T. Itoh, T. Suga, Self-excited force-sensing microcantilevers with piezoelectric thin<br />
films for dynamic scanning force microscopy, Sensors Actuat. A 54 (1996) 477–481.<br />
[11] C. Lee, T. Itoh, T. Suga, Self-excited piezoelectric PZT microcantilevers for dynamic SFM<br />
– with inherent sensing <strong>and</strong> actuating capabilities, Sensors Actuat. A 72 (1999) 179–188.<br />
[12] Y. Su, A. Brunnschweiler, A.G.R. Evans, G. Ensell, Piezoresistive silicon V-AFM cantilevers<br />
for high-speed imaging, Sensors Actuat. A 76 (1999) 139–144.<br />
[13] H. Takahashi, K. Ando, Y. Shirakawabe, Self-sensing piezoresistive cantilever <strong>and</strong> its<br />
magnetic force microscopy applications, Ultramicroscopy 91 (2002) 63–72.
REFERENCEs 25<br />
[14] J. Thaysen, A. Boisen, O. Hansen, S. Bouwstra, Atomic force microscopy probe with<br />
piezoresiostive read-out <strong>and</strong> a highly symmetrical Wheatstone bridge arrangement,<br />
Sensors Actuat. A 83 (2000) 47–53.<br />
[15] S. Akamine, R.C. Barrett, C.F. Quate, Improved atomic force microscope images using<br />
microcantilevers with sharp tips, Appl. Phys. Lett. 57 (3) (1990) 316–318.<br />
[16] T.R. Albrecht, S. Akamine, T.E. Carver, C.F. Quate, Microfabrication of cantilever styli<br />
for the atomic force microscope, J. Vac. Sci. Technol. A 8 (4) (1990) 3386–3396.<br />
[17] J. Brugger, R.A. Buser, N.F. de Rooij, Silicon cantilevers <strong>and</strong> tips for scanning force<br />
microscopy, Sensors Actuat. A 34 (1992) 193–200.<br />
[18] C. Liu, R. Gamble, Mass-producible monolithic silicon probes for scanning probe<br />
microscopes, Sensors Actuat. A 71 (1998) 233–237.<br />
[19] O. Wolter, T. Bayer, J. Greschner, Micromachined silicon sensors for scanning force<br />
microscopy, J. Vac. Sci. Technol. B 9 (2) (1990) 1353–1357.<br />
[20] S. Hosaka, K. Etoh, A. Kikukawa, H. Koyanagi, K. Itoh, 6.6 MHz silicon AFM cantilever<br />
for high-speed readout in AFM based recording, Microelectron. Eng. 46 (1999)<br />
109–112.<br />
[21] S. Habermehl, Stress relaxation in Si-rich silicon nitride thin films, J. Appl. Phys. 83 (9)<br />
(1998) 4672–4677.<br />
[22] K.E. Peterson, Silicon as a mechanical material, Proc. IEEE 70 (5) (1982) 420–457.<br />
[23] A. Khan, J. Philip, P. Hess, Young’s modulus of silicon nitride used in scanning force<br />
microscope cantilevers, J. Appl. Phys. 95 (4) (2004) 1667–1672.<br />
[24] J.J. Wortman, R.A. Evans, Young’s modulus, shear modulus, <strong>and</strong> Poisson’s ratio in<br />
silicon <strong>and</strong> germanium, J. Appl. Phys. 36 (1) (1965) 153–156.<br />
[25] T. Hantschel, S. Slesazeck, P. Niedermann, P. Eyben, W. V<strong>and</strong>ervorst, Integrating<br />
diamond pyramids into metal cantilevers <strong>and</strong> using them as electrical AFM probes,<br />
Microelectron. Eng. 57–58 (2001) 749–754.<br />
[26] K. Unno, T. Shibata, E. Makino, Micromachining of diamond probes for atomic force<br />
microscopy applications, Sensors Actuat. A 88 (2001) 247–255.<br />
[27] W. Kulisch, A. Malave, G. Lippold, W. Scholtz, C. Mihalcea, E. Oesterschulze,<br />
Fabrication of integrated diamond cantilevers with tips for SPM applications,<br />
Diamond Relat. Mater. 6 (1997) 906–911.<br />
[28] T. Akiyama, U. Staufer, N.F. De Rooij, L. Howald, L. Sc<strong>and</strong>ella, Lithographically<br />
defined polymer tips for quartz tuning fork based scanning probe microscopes,<br />
Microelectron. Eng. 58-58 (2001) 769–773.<br />
[29] R. Erl<strong>and</strong>sson, G. Hadziioannou, C.M. Mate, G.M. McClellamd, S. Chiang, Atomic<br />
scale friction between the muscovite mica cleavage plane <strong>and</strong> a tungsten tip, J. Chem.<br />
Phys. 89 (8) (1988) 5190–5192.<br />
[30] C.M. Mate, G.M. McClellamd, R. Erl<strong>and</strong>sson, S. Chiang, Atomic-scale friction of a<br />
tungsten tip on a graphite surface, Phys. Rev. Lett. 59 (17) (1987) 1942–1946.<br />
[31] Y. Miyahara, T. Fujii, S. Watanabe, A. Tonoli, S. Carabelli, H. Yamada, H. Bleuler, Lead<br />
zirconate titanate cantilever for non-contact atomic force microscopy, Appl. Surf. Sci.<br />
140 (1999) 428–431.<br />
[32] A.H. Sorenson, U. Hvid, M.W. Mortensen, K.A. Morch, Preparation of platinum/<br />
iridium scanning probe microscopy tips, Rev. Sci. Instrum. 70 (7) (1999) 3059–3067.<br />
[33] E.I. Givargizov, A.N. Stepanova, E.S. Mashkova, V.A. Molchanov, F. Shi, P. Hudek,<br />
I.W. Rangelow, Ultrasharp diamond coated silicon tips for scanning probe devices,<br />
Microelectron. Eng. 41/42 (1998) 499–502.<br />
[34] E.I. Givargizov, A.N. Stepanova, L.N. Obelenskaya, E.S. Mashkova, V.A. Molchanov,<br />
M.E. Givargizov, I.W. Rangelow, Whisker probes, Ultramicroscopy 82 (2000) 57–61.<br />
[35] S.S. Wong, E. Joselevich, A.T. Woolley, C.L. Cheung, C.M. Lieber, Covalently functionalized<br />
nanotubes as nanometre sized probes in chemistry <strong>and</strong> biology, Nature 394<br />
(1998) 52–55.
26 1. BAsIC PRINCIPLEs OF ATOMIC FORCE MICROsCOPy<br />
[36] C.L. Cheung, J.H. Hafner, T.W. Odom, K. Kim, C.M. Lieber, Growth <strong>and</strong> fabrication<br />
with single-walled carbon nanotube probe microscopy tips, Appl. Phys. Lett. 76 (21)<br />
(2000) 3136–3138.<br />
[37] H.J. Dai, J.H. Hafner, A.G. Rinzler, D.T. Colbert, R.E. Smalley, Nanotubes as nanoprobes<br />
in scanning probe microscopy, Nature 384 (6605) (1996) 147–150.<br />
[38] C.T. Gibson, S. Carnally, C.J. Roberts, Attachment of carbon nanotubes to atomic force<br />
microscope probes, Ultramicroscopy 107 (10–11) (2007) 1118–1122.<br />
[39] J.H. Hafner, C.L. Cheung, C.M. Lieber, Growth of nanotubes for probe microscopy<br />
tips, Nature 398 (6730) (1999) 761–762.<br />
[40] E. Bonnaccurso, G. Gillies, Revealing contamination on AFM cantilevers by microdrops<br />
<strong>and</strong> microbubbles, Langmuir 20 (2004) 11824–11827.<br />
[41] T. Thundat, X.-Y. Zheng, G.Y. Chen, S.L. Sharp, R.J. Warmack, L.J. Schowalter,<br />
Characterization of atomic force microscope tips by adhesion force measurements,<br />
Appl. Phys. Lett. 63 (15) (1993) 2150–2152.<br />
[42] L. Sirghi, O. Kylian, G. Ceccone, F. Rossi, Cleaning <strong>and</strong> hydrophobization of atomic<br />
force microscopy silicon probes, J. Phys. Chem. B 110 (2006) 25975–25981.<br />
[43] Y.-S. Lo, N.D. Huefner, W.S. Chan, P. Dryden, B. Hagenhoff, T.P. Beebe Jr., Organic<br />
<strong>and</strong> inorganic contamination on commercial AFM cantilevers, Langmuir 15 (1999)<br />
6522–6526.<br />
[44] S.W. King, R.J. Nemanich, R.F. Davis, Wet chemical processing of (0001)Si 6H-SiC<br />
hydrophobic <strong>and</strong> hydrophilic surfaces, J. Electrochem. Soc. 146 (5) (1999) 1910–<br />
1917.<br />
[45] D.A. Dobbs, R.G. Bergman, K.H. Theopold, Piranha solution explosion, Chem. Eng.<br />
News 68 (17) (1990) 2.<br />
[46] C.V. Erickson, Piranha solution explosions, Chem. Eng. News 68 (33) (1990) 2.<br />
[47] T. Arai, M. Tomitori, Removal of contamination <strong>and</strong> oxide layers from UHV-AFM tips,<br />
Appl. Phys. A 66 (1998) S319–S323.<br />
[48] K. Choi, T.-J. Eom, C. Lee, Comparison of the removal efficiency for organic contaminants<br />
on silicon wafers stored in plastic boxes between UV/O 3 <strong>and</strong> ECR oxygen<br />
plasma cleaning methods, Thin Solid Films 435 (2003) 227–231.<br />
[49] E. Bonaccurso, G. Gillies, Revealing contamination on AFM cantilevers by microdrops<br />
<strong>and</strong> microbubbles, Langmuir 20 (2004) 11824–11827.<br />
[50] R. Luginbuhl, A. Szuchmacher, M.D. Garrison, J.-B. Lhoest, R.M. Overney,<br />
B.D. Ratner, Comprehensive surface analysis of hydrophobically functionalized SFM<br />
tips, Ultramicroscopy 82 (2000) 171–179.<br />
[51] L. Montellius, J.O. Tegenfeldt, Direct observation of the tip shape in scanning probe<br />
microscopy, Appl. Phys. Lett. 62 (21) (1993) 2628–2630.<br />
[52] E.J. van Loenen, D. Dijkamp, A.J. Hoeven, J.M. Lenssink, J. Dieleman, Evidence for tip<br />
imaging in scanning tunneling microscopy, Appl. Phys. Lett. 56 (18) (1990) 1755–1757.<br />
[53] H.G. Hansma, R.L. Sinsheimer, J. Groppe, T.C. Bruice, V. Elings, M. Bezanilla,<br />
I.A. Mastrangelo, P.V.C. Hough, P.K. Hansma, Recent advances in atomic-force microscopy<br />
of DNA, Scanning 15 (5) (1993) 296–299.<br />
[54] H.G. Hansma, J.P. Clevel<strong>and</strong>, M. Radmacher, D.A. Walters, P.E. Hillner, M. Bezanilla,<br />
M. Fritz, D. Vie, H.G. Hansma, C.B. Prater, J. Massie, L. Fukunaga, J. Gurley, V. Elings,<br />
Tapping mode atomic force microscopy in liquids, Appl. Phys. Lett. 64 (13) (1994)<br />
1738–1740.<br />
[55] Q. Zhong, D. Inniss, K. Kjoller, V. Elings, Fractured polymer/silica fiber surface studied<br />
by tapping mode atomic force microscopy, Surf. Sci. Lett. (1993) L688–L692.<br />
[56] I. Schmitz, M. Schreiner, G. Friedbacher, M. Grasserbauer, Phase imaging as an extension<br />
to tapping mode AFM for the identification of material properties on humidity<br />
sensitive surfaces, Appl. Surf. Sci. 115 (1997) 190–198.<br />
[57] R.S. McLean, B.B. Sauer, Tapping mode AFM studies using phase detection for resolution<br />
of nanophases in segmented polyurethanes <strong>and</strong> other block copolymers,<br />
Macromolecules 30 (1997) 8314–8317.
REFERENCEs 27<br />
[58] J. Loos, The art of SPM: scanning probe microscopy in materials science, Adv. Mater.<br />
17 (2005) 1821–1833.<br />
[59] R. Lüthi, E. Meyer, L. Howald, H. Haefke, D. Asnselmetti, M. Dreier, M. Rüetschi,<br />
T. Bonner, R.M. Overney, J. Frommer, H.-J. Güntherodt, Progress in noncontact<br />
dynamic force microscopy, J. Vac. Sci. Technol. B 12 (3) (1994) 1673–1676.<br />
[60] M. Radmacher, M. Fritz, J.P. Clevel<strong>and</strong>, D.A. Walters, P.K. Hansma, Imaging adhesion<br />
forces <strong>and</strong> elasticity of lysozyme adsorbed on mica with the atomic force microscope,<br />
Langmuir 10 (1994) 3809–3814.<br />
[61] N. Chaiyut, T. Amornsakchai, S. Thanawan, Force volume imaging of defects in<br />
highly drawn high-density polyethylene, Polym. Test. 26 (2007) 396–401.<br />
[62] C. Rotsch, M. Radmacher, Mapping local electrostatic forces with the atomic force<br />
microscope, Langmuir 13 (1997) 2825–2832.<br />
[63] H. Hillborg, N. Tomczak, A. Olah, H. Schonherr, G.J. Vancso, Nanoscale hydrophobic<br />
recovery: a chemical force microscopy study of UV/ozone-treated cross-linked<br />
poly(dimethylsiloxan), Langmuir 20 (2004) 785–794.<br />
[64] H. Schonherr, C.L. Feng, N. Tomczak, G.J. Vancso, Compositional mapping of polymer<br />
surfaces by chemical force microscopy down to the nanometer scale: reactions in<br />
block copolymer microdomains, Macromol. Symp. 230 (2005) 149–157.<br />
[65] E. A-Hassan, W.F. Heinz, M.D. Antonik, N.P. D’Costa, S. Nageswaran,<br />
C.-A. Schoenenberger, J.H. Hoh, Relative microelastic mapping of living cells by<br />
atomic force microscopy, Biophy. J. 74 (3) (1998) 1564–1578.<br />
[66] M. Fritz, M. Radmacher, N. Peterson, H.E. Gaub, Visualization <strong>and</strong> identification of<br />
intracellular structures by force, modulation microscopy <strong>and</strong> drug induced degradation,<br />
J. Vac. Sci. Technol. B 12 (3) (1994) 1526–1529.<br />
[67] J.S. Jourdan, S.J. Cruchon-Dupeyrat, Y. Huan, P.K. Kuo, G.Y. Liu, Imaging nanoscopic<br />
elasticity of thin film materials by atomic force microscopy: effects of force modulation<br />
frequency <strong>and</strong> amplitude, Langmuir 15 (1999) 6495–6504.<br />
[68] J.A. Otero, J.M. Lena, P. Pradanos, F. Tejerina, A. Hern<strong>and</strong>ez, Characterisation of nanofiltration<br />
membranes. Structural analysis by the DSP model <strong>and</strong> microscopical techniques,<br />
J. Membr. Sci. 279 (2006) 410–417.<br />
[69] W.J. Price, P.K. Kuo, T.R. Lee, R. Colorado Jr., Z.C. Ying, G.Y. Liu, Probing the local<br />
structure <strong>and</strong> mechanical response of nanostructures using force modulation <strong>and</strong><br />
nanofabrication, Langmuir 21 (2005) 5422–5428.<br />
[70] W.J. Price, S.A. Leigh, S.M. Hsu, T.E. Patten, G. Liu, Measuring the size dependence of<br />
Young’s modulus using force modulation atomic force microscopy, J. Phys. Chem. A<br />
110 (2006) 1382–1388.<br />
[71] R.W. Carpick, D.F. Ogletree, M. Salmeron, Lateral stiffness: a new nanomechanical<br />
measurement for the determination of shear strengths with friction force microscopy,<br />
Appl. Phys. Lett. 70 (1997) 1548–1550.<br />
[72] C.T. Gibson, G.S. Watson, S. Myhra, Lateral force microscopy, Wear 213 (1997) 72–79.<br />
[73] M.A. Lantz, S.J. O’Shea, A.C.F. Hoole, M.E. Well<strong>and</strong>, Lateral stiffness of the tip <strong>and</strong> tipsample<br />
contact in frictional force microscopy, Appl. Phys. Lett. 70 (8) (1997) 970–972.<br />
[74] E. Tocha, H. Schonherr, G.J. Vancso, Quantitative nanotribology by AFM: a novel universal<br />
calibration platform, Langmuir 22 (2006) 2340–2350.<br />
[75] S.K. Kaliapan, B. Capella, Temperature dependent elastic–plastic behaviour of polystyrene<br />
studied using AFM force distance curves, Polymer 46 (2005) 11416–11423.<br />
[76] J. Domke, M. Radmach, Measuring the elastic properties of thin polymer films with<br />
the atomic force microscope, Langmuir 14 (1998) 3320–3325.<br />
[77] C. Plassard, E. Lesniewska, I. Pochard, A. Nonat, Investigation of the surface structure<br />
<strong>and</strong> elastic properties of calcium silicate hydrates at the nanoscale, Ultramicroscopy<br />
100 (2004) 331–338.<br />
[78] A. Vinckier, I. Heyvaert, A. D’Hoore, T. McKittrick, C. Van Haesendonck,<br />
Y. Engelboghs, L. Hellemans, Immobilizing <strong>and</strong> imaging microtubules by atomic force<br />
microscopy, Ultramicroscopy 57 (1995) 337–343.
28 1. BAsIC PRINCIPLEs OF ATOMIC FORCE MICROsCOPy<br />
[79] I. Dulinska, M. Targosz, W. Strojny, M. Lekka, P. Czuba, W. Balwierz, M. Szymonski,<br />
Stiffness of normal <strong>and</strong> pathological erythrocytes studied by means of atomic force<br />
microscopy, J. Biochem. Biophys. Meth. 66 (2006) 1–11.<br />
[80] A.L. Weisenhorn, P. Maivald, H.-J. Butt, P.K. Hansma, Measuring adhesion, attraction<br />
<strong>and</strong> repulsion between surfaces in liquids with an atomic force microscope, Phys. Rev.<br />
B 45 (19) (1992) 11226–11233.<br />
[81] C.J. Roberts, What can we learn from atomic force microscopy adhesion measurements<br />
with single drug particles?, Eur. J. Pharm. Sci. 24 (2005) 153–157.<br />
[82] M. Kapple, H.-J. Butt, The colloidal probe technique <strong>and</strong> its application to adhesion<br />
force measurements, Part. Part. Syst. Charact. 19 (2002) 129–143.<br />
[83] W.R. <strong>Bowen</strong>, N. <strong>Hilal</strong>, R.W. Lovitt, C.J. Wright, An atomic force microscopy study of<br />
the adhesion of a silica sphere to a silica surface – effects of surface cleaning, Colloids<br />
Surf. A: Physicochem Eng. Aspects 157 (1999) 117–125.<br />
[84] J.K. Eve, N. Patel, S.Y. Luk, S.J. Ebbens, C.J. Roberts, A study of single drug particle<br />
adhesion interactions using atomic force microscopy, Int. J. Pharm. 238 (2002) 17–27.<br />
[85] L. Xu, B.E. Logan, Analysis of bacterial adhesion using a gradient force analysis<br />
method <strong>and</strong> colloid probe microscopy, Langmuir 20 (2004) 8817–8822.<br />
[86] N.A. Burnham, D.D. Dominguez, R.L. Mowery, R.J. Colton, Probing the surface forces<br />
of monolayer films with an atomic force microscope, Phys. Rev. Lett. 64 (16) (1990)<br />
1931–1934.<br />
[87] M. Davies, A. Brindley, X. Chen, M. Marlow, S.W. Doughty, I. Shrubb, C.J. Roberts,<br />
Characterization of drug particle surface energetics <strong>and</strong> Young’s modulus by atomic<br />
force microscopy, Pharm. Res. 22 (7) (2005) 1158–1166.<br />
[88] G. Gillies, M. Kapple, H.-J. Butt, Surface <strong>and</strong> capillary forces encountered by zinc<br />
sulfide microspheres in aqueous electrolyte, Langmuir 21 (2005) 5882–5886.<br />
[89] P.G. Hartley, F. Grieser, P. Mulvaney, G.W. Stevens, Surface forces <strong>and</strong> deformation<br />
at the oil–water interface probed using AFM force measurement, Langmuir 15 (1999)<br />
7282–7289.<br />
[90] S. Allen, X. Chen, J. Davies, M.D. Davies, A.C. Dawkes, J.C. Edwards, C.J. Roberts,<br />
J. Sefton, S.J.B. Tendler, P.M. Williams, Detection of antigen–antibody binding events<br />
with the atomic force microscope, Biochemistry 36 (1997) 7457–7463.<br />
[91] R.B. Best, D.J. Brockwell, J.L. Toca-Herrerra, A.W. Blake, D.A. Smith, S.E. Radford,<br />
J. Clarke, Force mode atomic force microscopy as a tool for protein folding studies,<br />
Anal. Chim. Acta 479 (2003) 87–105.<br />
[92] J. Clarke, P.M. Williams, Unfolding induced by mechanical force, in: J. Buchner <strong>and</strong><br />
T. Kiefhaber (Ed.), Protein Folding H<strong>and</strong>book, Part 1, Wiley-VCH, 2005, pp. 1111–1142.<br />
[93] M. Rief, M. Gautel, F. Oesterhelt, J.M. Fern<strong>and</strong>ez, H.E. Gaub, Reversible unfolding of<br />
individual titin immunoglobulin domains by AFM, Science 276 (1997) 1109–1113.<br />
[94] M.J. Higgins, R. Proksch, J.E. Sader, M. Polcik, S. McEndoo, J.P. Clevel<strong>and</strong>, S.P. Jarvis,<br />
Noninvasive determination of optical lever sensitivity in atomic force microscopy,<br />
Rev. Sci. Instrum. 77 (2006) 013701.<br />
[95] J.P. Clevel<strong>and</strong>, S. Manne, D. Bocek, P.K. Hansma, A nondestructive method for determining<br />
the spring constant of cantilevers for scanning force microscopy, Rev. Sci.<br />
Instrum. 64 (2) (1993) 403–405.<br />
[96] J.E. Sader, Parallel beam approximation for v-shaped atomic force microscopy cantilevers,<br />
Rev. Sci. Instrum. 66 (9) (1995) 4583–4586.<br />
[97] C.T. Gibson, D.J. Johnson, C. Anderson, C. Abell, T. Rayment, Method to determine<br />
the spring constant of atomic force microscope cantilevers, Rev. Sci. Instrum. 75 (2)<br />
(2004) 565–567.<br />
[98] J.E. Sader, I. Larson, P. Mulvaney, L.R. White, Method for the calibration of atomic<br />
force microscopy cantilevers, Rev. Sci. Instrum. 66 (7) (1995) 3789–3798.<br />
[99] J.E. Sader, Frequency response of cantilever beams immersed in viscous fluids with<br />
applications to the atomic force microscope, J. Appl. Phys. 84 (1) (1998) 64–76.
REFERENCEs 2<br />
[100] J.E. Sader, J.W.N. Chon, P. Mulvaney, Calibration of rectangular atomic force microscopy<br />
cantilevers, Rev. Sci. Instrum. 70 (10) (1999) 3967–3969.<br />
[101] J. Ralston, I. Larson, M.W. Rutl<strong>and</strong>, A.A. Feiler, M. Kleijn, Atomic force microscopy<br />
<strong>and</strong> direct surface force measurements (IUPAC technical report), Pure Appl. Chem.<br />
77 (2005) 2149–2170.<br />
[102] L. Hutter, J. Bechhoefer, Calibration of atomic force microscope tips, Rev. Sci.<br />
Instrum. 64 (7) (1993) 1868–1873.<br />
[103] R. Levy, M. Maaloum, Measuring the spring constant of atomic force microscope cantilevers:<br />
thermal fluctuations <strong>and</strong> other methods, Nanotechnology 13 (2002) 33–37.<br />
[104] H.-J. Butt, M. Jaschke, Calculation of thermal noise in atomic force microscopy,<br />
Nanotechnology 6 (1995) 1–7.<br />
[105] J.L. Hutter, Comment on tilt of atomic force microscope cantilevers: effect on spring<br />
constant <strong>and</strong> adhesion measurements, Langmuir 21 (2005) 2630–2632.<br />
[106] R.W. Stark, T. Drobek, W.M. Heckl, Thermomechanical noise of a free v shaped cantilever<br />
for atomic force microscopy, Ultramicroscopy 86 (2001) 207–215.<br />
[107] D.A. Walters, J.P. Clevel<strong>and</strong>, N.H. Thomson, P.K. Hansma, M.A. Wendman,<br />
G. Gurley, V. Elings, Short cantilevers for atomic force microscopy, Rev. Sci. Instrum.<br />
67 (10) (1996) 3583–3590.<br />
[108] A. Torii, M. Sasaki, K. Hane, S. Okuma, A method for determining the spring constant<br />
of cantilevers for atomic force microscopy, Meas. Sci. Technol. 7 (1996) 179–184.<br />
[109] C.T. Gibson, G.S. Watson, S. Myra, Determination of the spring constants of probes<br />
for force microscopy/spectroscopy, Nanotechnology 7 (1996) 259–262.<br />
[110] S.M. Notley, S. Biggs, V.S.J. Craig, Calibration of colloid probe cantilevers using<br />
the dynamic viscous response of a confined liquid, Rev. Sci. Instrum. 74 (9) (2003)<br />
4026–4032.<br />
[111] V.S.J. Craig, C. Neto, In situ calibration of colloid probe cantilevers in force microscopy:<br />
hydrodynamic drag on a sphere approaching a wall, Langmuir 17 (2001)<br />
6018–6022.<br />
[112] N.A. Burnham, X. Chen, C.S. Hodges, G.A. Matei, E.J. Thoreson, C.J. Roberts,<br />
M.C. Davies, S.J.B. Tendler, Comparison of calibration methods for atomic-force<br />
microscopy cantilevers, Nanotechnology 14 (2003) 1–6.<br />
[113] T.J. Senden, W.A. Ducker, Experimental determination of spring constants in atomic<br />
force microscopy, Langmuir 10 (1994) 1003–1004.<br />
[114] J.M. Neumeister, W.A. Ducker, Lateral, normal <strong>and</strong> longitudinal spring constants of<br />
atomic force microscopy cantilevers, Rev. Sci. Instrum. 65-8 (1994) 2257–2531.<br />
[115] J.E. Sader, Susceptibility of atomic force microscope cantilevers to lateral forces, Rev.<br />
Sci. Instrum. 74 (4) (2003) 2438–2443.<br />
[116] W.A. Ducker, T.J. Senden, R.M. Pashley, Direct measurement of colloidal forces using<br />
an atomic force microscope, Nature 353 (1991) 239–241.<br />
[117] W.A. Ducker, T.J. Senden, Measurement of forces in liquids using a force microscope,<br />
Langmuir 8 (7) (1992) 1831–1836.<br />
[118] M. Götzinger, W. Peukert, Dispersive forces of particle–surface interactions: direct<br />
AFM measurements <strong>and</strong> modelling, Powder Technol. 130 (2003) 102–103.<br />
[119] X. Li-Chong, B.E. Logan, Interaction forces between colloids <strong>and</strong> protein-coated surfaces<br />
using an atomic force microscope, Environ. Sci. Technol. 39 (2005) 3592–3600.<br />
[120] G. Gillies, C.A. Prestidge, Colloid probe AFM investigation of the influence of<br />
crosslinking on the interaction behaviour <strong>and</strong> nano-rheology of colloidal droplets,<br />
Langmuir 21 (2005) 12342–12347.<br />
[121] M. Reitsma, V. Craig, S. Biggs, Elasto-plastic <strong>and</strong> visco-elastic deformations of a polymer<br />
sphere measured using colloid probe <strong>and</strong> scanning electron microscopy, Int.<br />
J. Adhesion Adhesives 20 (2000) 445–448.<br />
[122] L.B.R. Castro, M. Kapple, D.F.S. Petri, Adhesion forces between hybrid colloidal particles<br />
<strong>and</strong> concavilin A, Langmuir 22 (2006) 3757–3762.
30 1. BAsIC PRINCIPLEs OF ATOMIC FORCE MICROsCOPy<br />
[123] N. <strong>Hilal</strong>, W.R. <strong>Bowen</strong>, Atomic force microscopy study of the rejection of colloids by<br />
membrane pores, Desalination 150 (2002) 289–295.<br />
[124] W.R. <strong>Bowen</strong>, N. <strong>Hilal</strong>, R.W. Lovitt, C.J. Wright, Direct measurement of the force of<br />
adhesion of a single biological cell using an atomic force microscope, Colloids Surf.<br />
A: Physiochem. Eng. Aspects 136 (1998) 231–234.<br />
[125] X. Li, B.E. Logan, Analysis of bacterial adhesion using a gradient force analysis<br />
method <strong>and</strong> colloid probe atomic force microscopy, Langmuir 20 (2004) 8817–8822.<br />
[126] M.S. Barrow, W.S. <strong>Bowen</strong>, N. <strong>Hilal</strong>, A. Al-Hussany, P.R. Williams, R.L. Williams,<br />
C.J. Wright, A study of the tensile properties of liquids in confined spaces using an<br />
atomic force microscope, Proc. R. Soc. Lond. A 459 (2003) 2885–2908.
C H A P T E R<br />
2<br />
Measurement of Particle<br />
<strong>and</strong> Surface Interactions<br />
Using Force Microscopy<br />
<strong>Nidal</strong> <strong>Hilal</strong>, Daniel Johnson, W. <strong>Richard</strong><br />
<strong>Bowen</strong> <strong>and</strong> Paul M. Williams<br />
O U T L I N E<br />
2.1 Introduction 32<br />
2.2 Colloid Probes 32<br />
2.3 Interaction Forces 35<br />
2.3.1 van der Waals Forces 35<br />
2.3.2 Electrical Double Layer Forces 44<br />
2.3.3 DLVO Theory 53<br />
2.3.4 Solvation Forces 58<br />
2.3.5 Steric Interaction Forces 62<br />
2.3.6 Hydrophobic Interaction Forces 63<br />
2.3.7 Effect of Hydrodynamic Drag on AFM Force Measurements 66<br />
2.4 Adhesion Forces Measured by AFM 67<br />
2.4.1 Contact Mechanics <strong>and</strong> Adhesion 67<br />
2.5 Effect of Roughness on Measured Adhesion <strong>and</strong> Surface Forces 70<br />
Abbreviations <strong>and</strong> Symbols 70<br />
Greek Symbols 73<br />
References 74<br />
Atomic Force Microscopy in Process Engineering 31<br />
© 2009, Elsevier Ltd
32 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
2.1 IntRoduCtIon<br />
In Chapter 1, discussion was made of the application of the AFM to<br />
the measurement of forces. In this chapter, we will describe the use of the<br />
AFM to quantitatively measure surface forces arising from the interactions<br />
between particles <strong>and</strong> between particles <strong>and</strong> surfaces. These forces<br />
are of particular interest in the study of colloidal dispersions, where the<br />
strength of interactions between particles governs the properties of the<br />
dispersion overall.<br />
Most traditional methods used to study colloidal dispersions are<br />
ensemble techniques, where the interactions of a large number of particles<br />
are measured simultaneously. The advantage with the AFM is the ability<br />
to make measurements on the single-particle level <strong>and</strong> to measure forces,<br />
<strong>and</strong> hence interaction energies, with respect to intersurface distances.<br />
Depending upon the experimental set-up being employed, a number of<br />
different interactions may be measured, either separately or simultaneously,<br />
including long-range forces such as van der Waals <strong>and</strong> electrical<br />
double layer forces, hydrophobic interactions, solvation forces, steric<br />
interactions, hydrodynamic drag forces as well as adhesion. In this chapter,<br />
the forces that are likely to be encountered when measuring interactions<br />
between particles <strong>and</strong> between particles <strong>and</strong> surfaces whilst using<br />
the AFM are described, with examples of the measurements extant in the<br />
literature being included. It is hoped that this chapter will serve as a useful<br />
<strong>and</strong> practical guide to anyone who is undertaking such measurements.<br />
2.2 ColloId PRobES<br />
By replacing the pyramidal or conical probe usually present on an<br />
AFM cantilever with a microsphere, the geometry of interactions between<br />
the probe <strong>and</strong> surface can be greatly simplified, allowing the AFM to be<br />
used to probe surface forces, much akin to the surface force apparatus<br />
(SFA). Microparticles can also be used as probes, which will allow<br />
particle-to-particle adhesion forces to be measured. An example of a colloid<br />
probe is shown in Figure 2.1. Here, a scanning electron microscope (SEM)<br />
image shows a 5 �m diameter silica bead attached with glue close to the<br />
apex of a st<strong>and</strong>ard AFM microcantilever.<br />
As the size of a spherical colloid probe is increased, the potential area<br />
of contact will also increase in proportion. For two spheres in close proximity,<br />
the force as a function of distance D is:<br />
R R<br />
F( D)<br />
W( D)<br />
R R �<br />
1 2<br />
2π<br />
�<br />
⎛ ⎞<br />
⎜<br />
⎝⎜<br />
⎠⎟⎟<br />
1 2<br />
(2.1)
2.2 COLLOId PROBES 33<br />
FIGuRE 2.1 SEM image of a colloid probe created by the attachment of a glass sphere<br />
to the apex of an AFM microcantilever using an epoxy resin. The scale bar shown is 5 �m<br />
long, with the particle approximately 7.5 �m in diameter.<br />
where R 1 <strong>and</strong> R 2 are the radii of the two spheres <strong>and</strong> W(D) is the interaction<br />
energy per unit area as a function of distance [1, 2]. This relationship<br />
is referred to as the Derjaguin approximation. In the case of a sphere<br />
approaching a flat surface (or where one sphere is much greater in size than<br />
the other), this is further simplified to create:<br />
F( D) � 2πR1W( D)<br />
(2.2)<br />
Note that the forces of interaction are proportional to the radius or radii<br />
of the interacting spheres. For this reason, it is typical in colloidal probe<br />
measurements to normalise forces by dividing by the radius of curvature<br />
of the probe. This generally leads to forces being measured in units of the<br />
form of mN m �1 . It should be noted that one of the assumptions on which<br />
the Derjaguin approximation is based is that the range of the interaction<br />
forces is much less than the radii of the interacting spheres. For small<br />
spheres with radii in the nanometre range, this approximation is likely<br />
to be invalid, a situation that has been recently experimentally verified<br />
by carrying out measurements using AFM tips terminated with particles<br />
with radii of 10 <strong>and</strong> 20 nm [3]. Dividing the forces by the particle radii was<br />
insufficient to superimpose the force traces obtained with these particles.<br />
Colloidal probes are most commonly prepared by attaching the particle<br />
to the apex of the AFM cantilever with an appropriate glue using a
34 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
micromanipulator set-up, although where materials allow, sintering may be<br />
used, such as with polystyrene beads. Figure 2.2 shows a typical micromanipulator<br />
set-up, utilised in the preparation of colloid probes. This consists<br />
of a micromanipulator platform positioned underneath an optical microscope,<br />
<strong>and</strong> connected to an electronic control unit. Colloidal particles are<br />
placed upside down on a cleaned microscope slide, where they adhere via<br />
capillary forces. At one end of the slide is placed a thin smear of an appropriate<br />
adhesive. The AFM cantilever is allowed to come into contact with<br />
the glue, with care being taken to minimise the amount of glue present. Too<br />
much glue may cause problems with contamination. Any excess can be<br />
wiped off by scraping the cantilever carefully on a clean area of the slide.<br />
The stage is then moved until a suitable particle is located. The cantilever<br />
plus glue is then allowed to come into contact with the particle, removing it<br />
from the surface of the slide. The glue is allowed to set <strong>and</strong> the probe is then<br />
ready for use. An alternative method also used is to place a drop of glue on<br />
the end of the cantilever using a fine wire. Another wire is then used to pick<br />
up a particle using capillary adhesion <strong>and</strong> then place it on the glued end of<br />
the lever [4].<br />
C<br />
A<br />
D<br />
FIGuRE 2.2 Illustration of a typical micromanipulator set-up. It consists of a movable<br />
stage (A) mounted below an optical microscope (B). Inset into the top-left corner is a closeup<br />
of the stage. Movement of the stage can be controlled via an electronic control console<br />
(C), shown here on the left. Cantilever, particle interaction can also be monitored via a digital<br />
video camera (D) mounted on the microscope. The AFM cantilever is introduced to the<br />
underneath of a microscope slide mounted on the stage from the left, where it is allowed to<br />
come into contact with glue <strong>and</strong> particles attached to the slide. Fine control of the cantilever’s<br />
movement is attained via the manipulator joystick <strong>and</strong> vertical drive (E). On the monitor,<br />
a V-shaped AFM cantilever is visible. The picture was taken in the laboratory of the<br />
Department of Chemical <strong>and</strong> Environmental Engineering, University of Nottingham.<br />
A<br />
B<br />
E
2.3.1 van der Waals Forces<br />
2.3 IntERACtIon FoRCES<br />
In the nineteenth century, J. D. van der Waals derived an equation of<br />
state to account for deviations in the observed behaviour of gases from<br />
the ideal gas law, pV � nRT:<br />
⎛ 2<br />
n a ⎞<br />
⎜ VW ⎜<br />
⎜p<br />
� ( V � nbVW ) � nRT<br />
2<br />
⎝⎜<br />
V ⎠⎟<br />
(2.3)<br />
where p is the pressure, V is the volume of the container, n is the number<br />
of gas molecules in moles, R is the gas constant <strong>and</strong> T is the temperature.<br />
This equation contained two new terms to account for deviations from<br />
ideal behaviour, which can be determined experimentally <strong>and</strong> which vary<br />
between different types of molecule; a vw, which accounts for attractive<br />
forces between the molecules <strong>and</strong> b vw, which accounts for the volume of<br />
space occupied by the gas molecules <strong>and</strong> from which other molecules are<br />
excluded. This exclusion leads to repulsive forces at short ranges, due to the<br />
Pauli exclusion principle. Whilst there is a wide range of interactions that<br />
may occur between atoms, molecules <strong>and</strong> materials made from them, the<br />
term ‘van der Waals forces’ has been applied to a set of attractive forces that<br />
have their origin in interactions between dipoles, both permanent <strong>and</strong> temporarily<br />
induced, present in atoms <strong>and</strong> molecules. They can be divided<br />
into three types depending upon the precise nature of the interaction:<br />
dipole-dipole interactions (Keesom forces); dipole-induced dipole interactions<br />
(Debye forces) <strong>and</strong> interactions between dipoles induced on opposing<br />
atoms or molecules (London or dispersive forces), <strong>and</strong> thus all are of an<br />
electrostatic origin. The Keesom <strong>and</strong> Debye forces act between polar molecules.<br />
However, the London dispersive forces may arise between neutral<br />
atoms <strong>and</strong> are thus potentially present in all interactions between materials.<br />
All of these forces have interaction potentials of the form:<br />
w<br />
( r)<br />
2.3 INTERACTION FORCES 35<br />
CT<br />
( CK � CD � CL<br />
)<br />
� � �<br />
6 6<br />
r<br />
r<br />
(2.4)<br />
where w (r) is the interaction potential, C t is the constant of the interaction<br />
<strong>and</strong> r is the closest separation distance between molecules – subscripts<br />
K, D <strong>and</strong> L denote Keesom, Debye <strong>and</strong> London interactions respectively.<br />
When reading the literature, it is important to bear in mind which forces<br />
are being referred to by the term van der Waals forces. Some authors<br />
include all three of the types of dipole interactions mentioned here as<br />
van der Waals forces, whilst other authors specifically only mean the<br />
dispersion component of the interaction.
36 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
All van der Waals interactions decrease to the inverse sixth power<br />
of the separation distance. In effect, this means that these forces are not<br />
significant at ranges greater than the order of 100 nm <strong>and</strong> are unable<br />
to produce alignment effects in liquids at long ranges [2]. One simple<br />
approximation of the combined attractive van der Waals forces combined<br />
with the short-range electron shell repulsion is the Lennard-Jones<br />
potential [2, 5]:<br />
w<br />
( r)<br />
A B<br />
� � � � �<br />
r r r r<br />
4<br />
12 6<br />
� �<br />
� ⎛ ⎡ ⎞ ⎛ ⎞ ⎤<br />
⎢⎜<br />
⎜ ⎥<br />
⎢⎝⎜<br />
⎠⎟<br />
⎝⎜<br />
⎠⎟<br />
⎥<br />
⎣⎢<br />
⎦⎥<br />
6 12<br />
(2.5)<br />
where � is the depth of the potential energy well, � is the distance at<br />
which the potential energy is zero. In the simpler form on the right,<br />
A � 4�� 6 <strong>and</strong> B � 4�� 12 , which are the attractive <strong>and</strong> repulsive components<br />
of the Lennard-Jones potential respectively. Whilst the attractive<br />
component declines to the sixth power, the repulsive short-range forces<br />
decline to the twelfth power. This leads to a change in interaction potential<br />
with distance, as illustrated in Figure 2.3. At large distances, interaction<br />
forces are insignificant. On close approach between the molecules, attractive<br />
(negative sign) forces begin to dominate, with the potential reaching<br />
a minimum before repulsive forces from repulsion of opposing electron<br />
shells dominate.<br />
Keesom or orientational forces arise from the angle averaged interactions<br />
between permanent dipoles on opposing molecules. This gives rise<br />
to the following interaction free energy w (r):<br />
w<br />
( r)<br />
u u<br />
CK<br />
u1u2 � � � � for kT �<br />
2 6 6<br />
3( 4πε ε ) kTr r<br />
4πε<br />
ε r<br />
0<br />
1 2 2 2<br />
r<br />
0<br />
r<br />
3<br />
(2.6)<br />
where u 1 <strong>and</strong> u 2 are the dipole moments of the two molecules or atoms,<br />
ε 0 is the permittivity of free space (8.854 � 10 �12 C 2 J �1 m �1 ), ε r is the<br />
dielectric constant of the intervening medium, k is the Boltzmann constant<br />
(1.380 � 10 �23 J K �1 ) <strong>and</strong> T is the absolute temperature. The interaction<br />
between the dipoles increases the probability of an orientation<br />
between the dipoles, which leads to mutual attraction.<br />
The Debye, or induction, interaction is a result of permanent dipoles<br />
inducing temporary dipoles in opposing molecules <strong>and</strong> is the angle averaged<br />
interaction between such dipoles. The resulting free energy from<br />
such an interaction is:<br />
w<br />
( r)<br />
[ u 02 � u2<br />
] C<br />
� �<br />
� �<br />
( ) r r<br />
2 α �01<br />
4πε<br />
ε<br />
1 2<br />
0<br />
r<br />
2 6<br />
D<br />
6<br />
(2.7)
Interaction potential, w(r)<br />
0<br />
where � 01 <strong>and</strong> � 02 are the electronic polarisabilities of the two molecules.<br />
Both the Keesom <strong>and</strong> the Debye interactions involve polar molecules<br />
<strong>and</strong> are thus not always present, depending upon the molecules involved<br />
in the interaction. However, the dispersive, or London, component of<br />
the interaction described by London during the 1930s is always present.<br />
As two molecules come into close proximity, the repulsion between<br />
the negative charges in the electron shells causes the induction of temporary<br />
dipoles. As the molecules do not have to be polar <strong>and</strong> can be<br />
electrically neutral, this interaction can <strong>and</strong> does occur between any molecules<br />
within sufficient range of each other. The interaction free energy<br />
for the dispersion interaction between two molecules can be described as<br />
follows [6, 7]:<br />
w<br />
( r)<br />
2.3 INTERACTION FORCES 37<br />
3 �01�02 hv1v2 C<br />
� �<br />
� �<br />
2 6 2 ( 4πε<br />
) r ( v � v ) r<br />
0<br />
1 2<br />
w (r)<br />
Attractive<br />
Repulsive<br />
0.2 0.4 0.6 0.8 1<br />
Distance (nm)<br />
FIGuRE 2.3 Sketch of the Lennard-Jones potential for an interacting molecular pair<br />
described by equation (2.5). Also shown are the individual attractive <strong>and</strong> repulsive components.<br />
As the two molecules approach, attractive forces increase until repulsive forces due<br />
to the proximity of the electron shells of the molecules overcome the attraction.<br />
L<br />
6<br />
(2.8)<br />
where v 1 <strong>and</strong> v 2 are the orbiting frequencies of electrons, <strong>and</strong> h is Planck’s<br />
constant (6.626 � 10 � 34 m 2 kg s �1 ).
38 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
2.3.1.1 van der Waals Forces Between Bulk Materials<br />
For the useful estimation or measurement of van der Waals interaction<br />
forces between colloidal particles <strong>and</strong>/or surfaces, the theory outlined<br />
in the previous section needs to be extrapolated to describe the behaviour<br />
between materials rather than purely between individual atoms or<br />
molecules. A molecule at the surface of a particle in close proximity to<br />
another will interact with its neighbouring molecules, with molecules<br />
on the opposing particle as well as with the constituent molecules of the<br />
intervening medium. The summation of all the pair potentials interacting<br />
between macroscopic bodies results in forces that decay much more<br />
slowly with distance than is the case for single molecular interactions [2].<br />
The effect of the dispersion force was investigated theoretically by<br />
Hamaker [8] <strong>and</strong> de Boer [9]. For interactions between spherical particles,<br />
they used a pairwise summation of the interatomic dispersion energies <strong>and</strong><br />
demonstrated that although the range of the atomic forces was of the range<br />
of atomic dimensions, the sum of all of the dispersion energies resulted in<br />
an interaction range for colloidal bodies of the order of their dimensions.<br />
In other words, when scaled up to particles containing a great number of<br />
atoms, the range of the forces no longer decreases by the sixth power of<br />
the distance when separation is small compared to the size of the particles<br />
[10]. The coefficient of interaction used by Hamaker is now referred to as<br />
the Hamaker constant (A H). However, pairwise additivity does not hold<br />
especially if the interaction is occurring in a condensed medium, such as<br />
a liquid. This is because nearby atoms affect the forces acting between any<br />
interacting pair of molecules within a close vicinity [2].<br />
Lifshitz later described a macroscopic approach that completely<br />
avoided the problems associated with additivity, neglecting atomic structure,<br />
treating large bodies as continuous media with forces being derived<br />
in terms of bulk properties such as the dielectric constants <strong>and</strong> refractive<br />
indices [11]. In Table 2.1, the interaction energies <strong>and</strong> forces are listed for<br />
geometries of spherical particles interacting with other spheres <strong>and</strong> flat<br />
planes in terms of the Hamaker constant, A H, minimum separation distance<br />
<strong>and</strong> sphere radii. Here, the force is the negative differential of the<br />
potential with respect to the minimum separation:<br />
F<br />
dVA<br />
� �<br />
dD<br />
(2.9)<br />
The Hamaker constant contains elements describing all the material<br />
properties of the systems of interest. It can be summarised by the following<br />
formula:<br />
A � π C�<br />
�<br />
H<br />
2<br />
1 2<br />
(2.10)
TABLE 2.1 Energy <strong>and</strong> force expressions for different geometries.<br />
Two spheres<br />
R 1<br />
D<br />
Identical spheres<br />
R<br />
D<br />
Sphere – Plane<br />
R<br />
D<br />
Paraboloid – Plane<br />
R 2<br />
R<br />
2.3 INTERACTION FORCES 39<br />
V<br />
V<br />
V<br />
V<br />
A<br />
A<br />
A<br />
A<br />
AH<br />
=<br />
R R<br />
D R R<br />
−<br />
1 2<br />
6 +<br />
A R H<br />
=<br />
D<br />
−<br />
12<br />
A R H<br />
=<br />
D<br />
−<br />
6<br />
A<br />
=<br />
l<br />
D l<br />
−<br />
12<br />
1 2<br />
A R R<br />
H<br />
F =<br />
D R R<br />
−<br />
6 +<br />
2<br />
1 2<br />
A R H<br />
F =<br />
D<br />
−<br />
12 2<br />
A R H<br />
F =<br />
D<br />
−<br />
6 2<br />
1 2<br />
where � 1 <strong>and</strong> � 2 are the number of atoms per unit volume in the two<br />
interacting materials.<br />
There are some general properties of van der Waals interactions<br />
between macroscopic bodies, which are worthy of note. First, interactions<br />
between two bodies across vacuum are always attractive, as are all interactions<br />
between two bodies of identical composition across a medium.<br />
However, for two bodies of dissimilar materials, the net interaction may<br />
be either attractive or repulsive, depending upon the particular setup.<br />
Whilst all van der Waals interactions per se are attractive, if one of<br />
the bodies has a greater attraction for the intervening medium than for<br />
the opposing body, then this will result in a net repulsion between the<br />
two bodies [12]. Whether such an interaction is likely to be attractive or<br />
repulsive can be assessed by comparison of the dielectric constants of the<br />
materials involved. If the dielectric constant of the intervening medium<br />
is between that of the materials of the two bodies, then the net forces<br />
between the two bodies will be repulsive. If it matches the dielectric constant<br />
of either of the interacting bodies, then the van der Waals force will<br />
effectively vanish [13].<br />
H xy<br />
2<br />
2<br />
z<br />
2<br />
xy<br />
2<br />
z<br />
A j<br />
H<br />
F =<br />
D l<br />
−<br />
12 2
40 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
2.3.1.2 Retardation of van der Waals Forces<br />
As the distance between interacting atoms increases, the time for the<br />
electric field of one atom to interact increases, <strong>and</strong> for a large enough distance<br />
will become comparable with the time over which the dipole itself<br />
fluctuates, leading to fluctuations in the interacting dipoles becoming out<br />
of step. This can lead to the interaction becoming less favourable, causing<br />
the strength of the interaction to decrease with the inverse seventh<br />
power of the separation distance rather than to the inverse sixth power<br />
[2, 14]. Because of this mechanism, it is only the London dispersion interactions<br />
that are affected by these retardation affects <strong>and</strong> not the Debye or<br />
Keesom interactions.<br />
2.3.1.3 Calculation of Hamaker Constants<br />
Looking at Table 2.1, it can be seen that if the Hamaker constant is<br />
known, it is possible to calculate the interaction energy between surfaces,<br />
provided that particle radii <strong>and</strong> distances of separation are available. The<br />
Hamaker constant though is not an easily obtained value.<br />
Lifshitz theory [11] can be used to calculate the Hamaker constant,<br />
but the calculations require complete knowledge of the dielectric spectra<br />
over the entire frequency range for all of the individual materials comprising<br />
the system. This type of data is not available for most substances,<br />
so another method is required for which data is more widely available.<br />
Ninham <strong>and</strong> Parsegian [15] considered the construction of the dielectric<br />
spectra for all frequencies <strong>and</strong> concluded that not all parts of the<br />
frequency range are equally important. They found that the ultraviolet<br />
absorption regime is the most important of all the contributions to the<br />
frequency sum. Hough <strong>and</strong> White [16] also emphasised the importance<br />
of the ultraviolet absorption peak. To obtain the parameters for the UV<br />
peak, refractive index data measured over a range of wavelengths, �,<br />
usually in the visible part of the spectrum, can be used. The following<br />
equation was then constructed [16]:<br />
2 2 ω<br />
no ( ω) � 1 � ( no ( ω)<br />
� 1)<br />
� C<br />
2<br />
ω<br />
2<br />
UV<br />
UV<br />
(2.11)<br />
π<br />
where ω�<br />
λ<br />
2 c , no(�) is the refractive index at a given frequency <strong>and</strong> c is<br />
the speed of light in a vacuum.<br />
2<br />
2<br />
Thus, if a graph of ( no � 1)<br />
is plotted versus ( no �1) �2 , a straight line<br />
2<br />
of slope 1/ωUV<br />
<strong>and</strong> intercept CUV will be obtained. This method of analysis<br />
is called the ‘Cauchy Plot’.
2.3 INTERACTION FORCES 41<br />
Horn <strong>and</strong> Israelachvili [17] presented an expression for the Hamaker<br />
constant, which is as follows:<br />
AH � A � Aξ � A<br />
o ξ�<br />
131 1<br />
(2.12)<br />
where the two terms in this equation may be found using the Cauchy<br />
Plot data from:<br />
where<br />
A<br />
ξ<br />
�1<br />
3<br />
⎛εr<br />
( 0) �ε<br />
r ( 0)<br />
⎞<br />
⎜ 1 2<br />
Aξ�kT⎜ o 4<br />
⎜<br />
⎝⎜<br />
εr<br />
( 0) �εr<br />
( 0)<br />
⎠⎟<br />
n<br />
o<br />
2 2<br />
o o<br />
n n<br />
�<br />
�<br />
2<br />
1 3<br />
2 2<br />
o o o<br />
∆n �n �n<br />
1 3<br />
ω1<br />
2 ω3<br />
2<br />
X � ( n o �1) � ( n o �1)<br />
1 3<br />
ω<br />
ω<br />
3<br />
1 2<br />
2 2<br />
o o o o<br />
3ℏ<br />
ωω 1 3 Xn�2XΔnn�Δn ( 3�2Y) �<br />
7/ 4<br />
64no<br />
⎡<br />
1/ 2<br />
/ ⎤<br />
⎢ 2<br />
2<br />
( Y � Y �1)<br />
� ( Y � Y � 1)<br />
⎥<br />
⎣<br />
⎢<br />
⎦<br />
⎥<br />
1 ⎡ ω1<br />
2 ω ⎤<br />
Y � ⎢<br />
3 2<br />
( n o �1) � ( n o �1)<br />
⎥<br />
1 3<br />
4 n ⎢<br />
o ⎣ω<br />
3 ω ⎥<br />
1 ⎦<br />
1<br />
(2.15)<br />
(2.16)<br />
(2.17)<br />
(2.18)<br />
<strong>and</strong> ε ( 0) ri is the dielectric constant of component i, � is Planck’s constant<br />
divided by 2�, k is Boltzmann’s constant, n is the refractive index of com-<br />
oi<br />
ponent i (calculated from C UV �n o �<br />
2<br />
1), T is the absolute temperature,<br />
�i is the ultraviolet characteristic adsorption frequency of component<br />
i (i.e. � �UVi), subscript 1 refers to the material <strong>and</strong> subscript 3 to the dispersion<br />
medium.<br />
Therefore, if the dielectric constant <strong>and</strong> refractive index versus wavelength<br />
data are known for each component, i, the Hamaker constant may<br />
be approximated by use of a Cauchy Plot <strong>and</strong> the above equations.<br />
2<br />
1 2 3<br />
(2.13)<br />
(2.14)
42 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
Equation (2.12) is the expression for the non-retarded Hamaker constant.<br />
When comparing Hamaker constant data with literature values,<br />
this equation should be used, as the values given in the literature are for<br />
non-retarded Hamaker constants. However, in calculations where particles<br />
in an electrolyte solution are considered, <strong>and</strong> for some cases at large<br />
interparticle separations, the effects of retardation [18] <strong>and</strong> screening [19]<br />
on the Hamaker constant calculation need to be taken into account. This<br />
can be done by modifying equation (2.12) to:<br />
where<br />
A 131 �A ( 1�2κD) e +<br />
A F( H)<br />
ξ<br />
o<br />
�2κD<br />
F H<br />
H<br />
⎡<br />
( ) � ⎢ ⎛ π ⎞<br />
1�⎜<br />
⎢ ⎝⎜<br />
4 2 ⎠⎟<br />
⎣⎢<br />
3/ 2<br />
�2/<br />
3<br />
D<br />
H �no ( n o �n<br />
3 1 o )<br />
3 c<br />
/ w w<br />
⎤<br />
⎥<br />
⎦⎥<br />
2 2 1 2 1 3<br />
(2.19)<br />
(2.20)<br />
(2.21)<br />
2.3.1.4 Calculation of van der Waals Forces from<br />
Force Distance Curves<br />
Measurement of the van der Waals interactions between different<br />
materials in different media is possible using the AFM. By changing<br />
the sharp probe with a colloidal particle, a wide range of different<br />
interactions can be measured. However, much care must be taken with<br />
the design of experiments. In many situations, forces other than van<br />
der Waals interactions may be present, such as double layer interactions<br />
<strong>and</strong> hydrodynamic effects, which may make it difficult or impossible<br />
to extract accurate estimations of van der Waals forces or Hamaker<br />
constants. In addition, any calculations made must take into account the<br />
probe sample interaction geometry. In the case of a spherical particle versus<br />
a plane surface or other spherical particles, this is relatively straightforward.<br />
However, for an irregularly shaped particle or one exhibiting<br />
significant surface roughness, then this is not a straightforward matter.<br />
This subject is dealt with in more detail in Section 2.5 of this chapter.<br />
There are several ways in which Hamaker constants may be estimated<br />
from AFM force measurements. First, one of the power laws, appropriate<br />
for the interaction geometry used in the experiment, from Table 2.1 may<br />
be fitted to the part of the force distance measurement obtained when<br />
the probe is approaching the surface just prior to any jump-to-contact.<br />
During the jump-to-contact event, the cantilever system is out of equilibrium<br />
<strong>and</strong> consequently, the power laws may not be fitted to this part<br />
ξ<br />
�1
of the force curve. Researchers have overcome this problem by the use<br />
of magnetically actuated cantilevers operating under a feedback mechanism<br />
to maintain the stability of the system [20].<br />
Butt et al., in a comprehensive review of force measurement techniques<br />
[21], describe a method for calculating the Hamaker constant from both the<br />
jump-in distance <strong>and</strong> the deflection of the cantilever due to the jump-in. This<br />
requires knowledge of the radius of curvature of the probe tip (or spherical<br />
particle replacing the probe) along with the effective stiffness of both the<br />
cantilever <strong>and</strong> total system. For a sphere approaching a plane surface,<br />
A<br />
H<br />
2.3 INTERACTION FORCES 43<br />
3<br />
jtc<br />
3<br />
kc<br />
⎟ keff<br />
2<br />
3k D ⎛<br />
eff jtc 2k ⎞<br />
⎜ 3k x<br />
C eff 24<br />
� � ⎜ �<br />
Rt<br />
⎝⎜<br />
keff<br />
⎠ Rt<br />
Rt<br />
(2.22)<br />
where D jtc is the jump-in distance, x jtc is the cantilever deflection due to<br />
the jump-in, R t is the radius of curvature of the probe tip, k c is the spring<br />
constant of the cantilever <strong>and</strong> k eff is the effective spring constant of the<br />
cantilever-sample system. The effective stiffness can be calculated from<br />
considering the contribution to the stiffness of the system of the cantilever<br />
<strong>and</strong> sample surface interacting in series:<br />
1 1 1<br />
� �<br />
keff kc ks<br />
(2.23)<br />
where k s is the stiffness of the sample, assuming negligible deformation<br />
of the probe. For hard samples or soft cantilevers, k eff � k c. Das <strong>and</strong> colleagues<br />
[22] used a similar approach to measure the Hamaker constant<br />
between silicon nitride AFM probes <strong>and</strong> a number of surfaces from the<br />
jump-in distance using the following relationship:<br />
A<br />
H<br />
k D<br />
�<br />
R<br />
24⎛<br />
⎜<br />
27⎝⎜<br />
c jtc<br />
t<br />
⎞<br />
⎠⎟<br />
(2.24)<br />
Measurements were carried out on a SiO 2 surface, freshly cleaved<br />
mica <strong>and</strong> on a silver metal film in air under ambient conditions. Values<br />
obtained were of the same magnitude, but not identical, to literature values<br />
obtained from Lifshitz theory <strong>and</strong> measurements with the surface<br />
forces apparatus (SFA). One possible reason for this divergence may be<br />
the presence of a contaminating water layer present on most surfaces<br />
under ambient conditions.<br />
The Hamaker constant for an interaction may also be calculated from the<br />
work of adhesion, W a, between the two bodies. The work of adhesion may<br />
be obtained from the adhesion as measured during the retract cycle of a force<br />
distance measurement <strong>and</strong> then applying the appropriate contact mechanics
44 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
model. These models will be described in more detail later in this chapter.<br />
The particular relationship between W a <strong>and</strong> A H depends again upon the particular<br />
geometry of the interaction. For a sphere–plane interaction [21],<br />
AH � 6D0 Wa<br />
(2.25)<br />
This is essentially the same as the force laws found in Table 2.1, replacing<br />
the minimum separation distance, D, with an interatomic spacing<br />
value. In turn, W a may be related to the interfacial energies by the following<br />
relationship:<br />
Wa � � � � � �<br />
13 23 12<br />
(2.26)<br />
where � is the interfacial energy, subscripts 1 <strong>and</strong> 2 denote the two solid<br />
bodies (in this case, the probe <strong>and</strong> the surface) <strong>and</strong> 3 denotes the intervening<br />
medium. With this method, there is great potential for error. Care<br />
must be taken when calculating surface forces from adhesion measurements.<br />
Interactions not present at long range may be present when contact<br />
is made, such as solvation forces as well as the effects of roughness<br />
<strong>and</strong> deformations. In one study [23], both long-range forces <strong>and</strong> adhesion<br />
measurements were observed between latex spheres <strong>and</strong> glass surfaces<br />
in aqueous solution using the colloidal probe technique. Adhesion values<br />
were estimated on the basis of the long-range interaction forces. These<br />
calculated adhesion values were approximately 20–30 times greater than<br />
the adhesion values actually measured.<br />
2.3.2 Electrical double layer Forces<br />
As stated above, van der Waals interactions between identical particles<br />
are always attractive. If this was the only force present between colloidal<br />
particles in solution, then dispersions would be unstable due to aggregation,<br />
leading to the formation of a precipitate. Fortunately, this is not the<br />
case as particles in water or any liquid of high dielectric constant usually<br />
possess charges on their surfaces. Repulsion between identically charged<br />
particles is long range in character <strong>and</strong> is often sufficient to overcome the<br />
aggregating effects of attractive van der Waals interactions.<br />
2.3.2.1 The Electrical Double Layer<br />
From observations of colloidal systems, it can be concluded that particles<br />
dispersed in water or any liquid with a high dielectric constant will<br />
usually develop a surface charge. The charging of a surface in a liquid<br />
can be brought about by one of two charging mechanisms [2]:<br />
1. By the ionisation or dissociation of surface groups, which leaves<br />
behind a charged surface (e.g. the dissociation of protons from carboxylic<br />
acid groups, which leaves behind a negatively charged surface).
2.3 INTERACTION FORCES 45<br />
2. By the adsorption (binding) of ions from solution onto a previously<br />
uncharged surface. The adsorption of ions from solution can also<br />
occur onto oppositely charged sites, also known as ion exchange.<br />
Since the system as a whole is electrically neutral, the dispersing<br />
medium must contain an equivalent charge of the opposite sign. These<br />
charges are carried by ions, i.e., by an excess of ions of one sign on the<br />
particle surface <strong>and</strong> an excess of ions of the opposite sign in the solution.<br />
Hence, if we consider an individual particle immersed in the liquid, it is<br />
surrounded by an electrical double layer. One part of this double layer is<br />
formed by the charge of the surface of the particles. Another part of the<br />
electrical double layer is formed by the excess of oppositely charged ions<br />
in the solution. As a result of their thermal motion, the electrical charge<br />
carried by this layer extends over a certain distance from the particle surface<br />
<strong>and</strong> dies out gradually with increasing distance (diffuse layer) into<br />
the bulk liquid phase.<br />
2.3.2.2 Distribution of Electrical Charge <strong>and</strong> Potential in the<br />
Double Layer<br />
The first approximate theory for the electrical double layer was given by<br />
Gouy, Chapman, Debye <strong>and</strong> Hückel [24]. In this theory, the average charge<br />
distribution <strong>and</strong> the corresponding electrical potential function have been<br />
related on the basis of the Poisson–Boltzmann equation (PBE) [18]:<br />
∇ 2 �1 0 �zie�<br />
� � ni zie exp<br />
ε ε<br />
k T<br />
⎛ ⎞<br />
⎜ ∑ ⎜<br />
⎝⎜<br />
⎠⎟<br />
0<br />
r<br />
i<br />
B<br />
(2.27)<br />
where � is the electrical potential, n i 0 is the number density of ions of<br />
valency z i, k B is the Boltzmann constant, T is the absolute temperature,<br />
�ε 0 is the permittivity of vacuum, ε r is the dielectric constant of the background<br />
solvent <strong>and</strong> e is the elementary charge.<br />
The above PBE has been deduced using a number of simplifying<br />
assumptions that the electrolyte is an ideal solution with uniform<br />
dielectric properties, the ions are point charges <strong>and</strong> the potential of<br />
mean force <strong>and</strong> the average electrostatic potential are identical. Besides,<br />
the PBE is only applicable to the system with a symmetrical electrolyte<br />
or a mixture of electrolytes of the same valency type. According to this<br />
theory, the average charge density at a given point can be calculated<br />
from the average value of the electrical potential at the same point with<br />
Boltzmann’s theorem. The electrical potential distribution can be related<br />
to the charge density with the aid of Poisson’s equation. As a matter of<br />
fact, the Gouy–Chapman theory has a rather serious defect, which is<br />
mainly a consequence of the neglect of the finite dimensions of the ions.
46 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
In dilute solutions, where the extension of the diffuse layer is considerable,<br />
this neglect is to some degree permissible; but in more concentrated<br />
electrolyte solutions, the picture in terms of the Gouy–Chapman model<br />
becomes incorrect in some essential details.<br />
Stern [25] modified the Gouy–Chapman model by taking into consideration<br />
the finite size of real ions, underlying the double layer theory for<br />
a solid wall by dividing the charges in liquid into two parts. One part is<br />
considered as a layer of ions adsorbed to the wall, <strong>and</strong> is represented in<br />
the theory by a surface charge concentrated in a plane at a small distance,<br />
d, from the surface charge on the wall, also known as the outer Helmholtz<br />
plane (OHP), as shown in Figure 2.4. The second part of the liquid charge<br />
is then taken to be a diffuse space charge, as in the old theory, extending<br />
from the OHP at x � d to infinity, where the PBE can be applied.<br />
The method with which the distance to the OHP is calculated depends<br />
on the type of model used for describing the compact region. For an<br />
oxide surface, such as typically found on the surface of silica, a triple<br />
layer model such as the Gouy–Chapman–Grahame–Stern model [26] is<br />
often used to describe the compact region; see Figure 2.4(a). This model<br />
allows for a plane of adsorbed ions (partially dehydrated) on the particle<br />
surface (the centres of which form the locus for the inner Helmholtz<br />
plane (IHP)) followed by a plane occurring at the distance of closest<br />
approach of the hydrated counterions (the OHP). This is the mechanism<br />
by which the high surface charge on the oxide is reconciled with the quite<br />
low diffuse double layer potentials (zeta potentials). For other types of<br />
–<br />
–<br />
–<br />
Particle<br />
surface<br />
–<br />
–<br />
–<br />
ε r1<br />
IHP<br />
+<br />
+<br />
+<br />
ε r2<br />
OHP<br />
+<br />
Aqueous<br />
solution<br />
x = 0 i d<br />
ψ = ψo ψi ψd<br />
σ = σo σi σd A B<br />
–<br />
–<br />
–<br />
Particle<br />
surface<br />
–<br />
–<br />
–<br />
x = 0<br />
ψ = ψo<br />
σ = σ o<br />
εr 1<br />
OHP<br />
+<br />
+<br />
+<br />
d<br />
ψd<br />
Aqueous<br />
solution<br />
FIGuRE 2.4 Models for compact part of the double layer: (A) Gouy–Chapman–<br />
Grahame–Stern (triple layer) model <strong>and</strong> (B) modified Gouy–Chapman model.<br />
σ d
surfaces such as proteins, where there are few or no adsorbed ions at all,<br />
the modified Gouy–Chapman model [26], where the OHP is located at<br />
the plane of closest approach of the hydrated counterions, is probably<br />
more appropriate; see Figure 2.4(b). The distance to the OHP can be calculated<br />
from the knowledge of the ionic crystal <strong>and</strong> hydrated ionic radii.<br />
The non-linear PBE is used to calculate the potential distribution<br />
inside the diffusive part of the electric double layer between two surfaces<br />
[18, 26]. According to the non-linear PBE, the aqueous solution is<br />
defined by its static dielectric constant only. The surface charge is usually<br />
taken as averaged over the surface, <strong>and</strong> the discrete nature of ions is not<br />
considered.<br />
In order to calculate the potential distribution around a particle, not<br />
only is the PBE needed, but the boundary conditions also have to be<br />
specified. A choice of boundary conditions is available at the particle surface.<br />
It is important to choose physically meaningful conditions at the<br />
particle surfaces, which depend on the colloidal material being considered<br />
(see next section).<br />
2.3.2.3 Interaction Forces Between Double Layers<br />
Analytical Solutions When two like-charged particles approach each<br />
other, their electrical double layers will begin to overlap, resulting in a<br />
repulsive force that will oppose further approach. For very dilute systems<br />
where just two particles can be considered in the interaction, it is possible<br />
to obtain analytical expressions for the calculation of the repulsive interaction<br />
energy between two spherical particles on the basis of the interaction<br />
energy equations derived for infinite flat plates of the same material<br />
with either the Derjaguin approximation [1] or the linear superposition<br />
approximation (LSA) [27], as shown below.<br />
V<br />
2.3 INTERACTION FORCES 47<br />
R �<br />
128 1 2<br />
1 2<br />
πa<br />
a n�kT � �<br />
( a � a ) �<br />
2 1 2<br />
exp( �κD)<br />
(2.28)<br />
where D is the surface-surface separation between the particles; a 1 <strong>and</strong> a 2<br />
denote the radii of particles 1 <strong>and</strong> 2; � is the Debye–Hückel reciprocal<br />
length; n o is the bulk density of ions <strong>and</strong> � is the reduced surface potential,<br />
which can be expressed as:<br />
�<br />
i<br />
⎛ze�i<br />
⎞<br />
� tanh⎜ ⎝⎜<br />
4kT<br />
⎠⎟⎟<br />
(2.29)<br />
The above equation is valid only when both the conditions �a � 5 <strong>and</strong><br />
D � � a are satisfied. There are many other expressions available on the<br />
basis of various assumptions for the sphere-sphere double layer interaction
48 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
energy. For further information, readers are referred to the literature<br />
[44–49]. In general, the LSA method yields the correct interaction at large<br />
separations for all surface potentials <strong>and</strong> particle sizes; Derjaguin’s integration<br />
gives accurate results for large particles at short distances <strong>and</strong> the<br />
McCartney <strong>and</strong> Levine formulation [33] is a good approximation at all<br />
separations except for small potentials. It should be noted that although<br />
the first two methods themselves place no restriction on the potentials,<br />
the resulting expressions often do because of the difficulty in solving the<br />
PBE. Therefore, care must be taken in choosing the correct expression.<br />
Numerical Solutions The previous analytical solutions only apply for<br />
a limited range of ionic strength, size <strong>and</strong> separation distance. This is<br />
mainly due to the fact that, for most cases, no exact analytical solution of<br />
the PBE is available. These limitations may be removed by considering<br />
numerical solutions of the problem being investigated.<br />
Before the advent of modern computers, a few numerical solutions for<br />
the PBE around a sphere were worked out [34]. A more extensive tabulation<br />
was compiled by Hoskin [35] using electronic computing techniques.<br />
Loeb et al. [36] then compiled a much more comprehensive set of tables,<br />
covering a wide variety of electrolyte types, concentrations <strong>and</strong> surface<br />
potentials. More recently, the PBE has been solved by various methods<br />
for unconfined [37–45] <strong>and</strong> confined [39, 46, 47] particles. These simulations<br />
can be multidimensional, <strong>and</strong> include Runge–Kutta, finite-element,<br />
finite difference <strong>and</strong> substitution methods. The case of two interacting<br />
particles isolated in a dielectric medium is usually considered as a test<br />
case for any new solution method as this system has been solved previously<br />
by many authors (e.g. see [38, 40–44]).<br />
As mentioned in Section 2.3.2.2, the PBE needs to be solved with<br />
respect to a set of boundary conditions. As an example, consider the case<br />
of an isolated sphere. The charge distribution of the counterions away<br />
from the particle is described by the non-linear PBE in spherical coordinates<br />
as:<br />
2<br />
d � 2 d�<br />
2n<br />
ze ez�<br />
� � sinh<br />
2<br />
dr r dr ε ε kT<br />
⎛ ⎞<br />
⎜<br />
⎝⎜<br />
⎠⎟<br />
0<br />
o<br />
r<br />
(2.30)<br />
Equation (2.30) cannot be solved analytically; so in order to solve the<br />
PBE numerically, two boundary conditions will be required. The condition<br />
that is used at the outer boundary is that of electroneutrality, given as:<br />
d�<br />
dr<br />
r�∞<br />
�0<br />
(2.31)
2.3 INTERACTION FORCES 49<br />
A choice of boundary conditions is available at the particle surface. It<br />
is important to choose physically meaningful conditions at the particle<br />
surface, which may depend on the colloidal material being considered.<br />
For metal sols in a solution, a constant surface potential boundary condition<br />
is appropriate, given by:<br />
� o r� �<br />
α<br />
constant<br />
(2.32)<br />
where � is the particle radius plus the distance to the OHP (� a � d)<br />
(see Figure 2.4).<br />
A constant surface charge boundary condition may be appropriate<br />
when the surface charge is caused by crystal lattice defects, such as in<br />
clay minerals.<br />
� � ε ε �<br />
o d r o<br />
d<br />
�� �� � constant<br />
dr<br />
r�α<br />
(2.33)<br />
A boundary condition where the zeta potential is held constant is also<br />
possible [37, 48].<br />
� d r=α<br />
� ζ � constant<br />
(2.34)<br />
In the case of biomaterials <strong>and</strong> oxide surfaces, the charge can be generated<br />
by surface dissociation reactions, which are influenced by the solution<br />
conditions. This can be described by a boundary condition known as<br />
charge regulation [49].<br />
� � f ( � ) ≠ constant<br />
o o<br />
(2.35)<br />
As an example of this, consider the protein bovine serum albumin<br />
(BSA). The protein is made up of a number of different types of amino<br />
acids. Only certain amino acids will take part in the ionisation reactions,<br />
which will generate a charge on the protein surface. The development of<br />
a charge regulation model for BSA requires the number of these chargegenerating<br />
amino acids to be known. This data is available in the literature<br />
from the amino acid sequence of the protein [50] or from titration<br />
data [51]. The relevant equilibria reactions are illustrated by:<br />
1<br />
� �<br />
for aspartic or glutamic acid � COOH �� �� �� ��<br />
� COO � H (2.36)<br />
2<br />
for lysine � NH �� �� �� ��<br />
� NH � H<br />
K<br />
� K<br />
�<br />
3 2<br />
(2.37)
50 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
Considering equation (2.36) as an example, the equilibrium constant<br />
for the reaction may be written as:<br />
K<br />
1 �<br />
ˆ<br />
�<br />
s<br />
[ COO ][ H ]<br />
[ COOH]<br />
(2.38)<br />
where [H � ] s is the hydrogen ion concentration at the BSA surface, which<br />
can be determined from:<br />
� + ⎛�ze<br />
o<br />
[ H ] s � [ H ] bulk exp⎜<br />
� ⎞<br />
⎝⎜<br />
kT ⎠⎟<br />
(2.39)<br />
with the bulk hydrogen ion concentration being found from the pH of<br />
the dispersion.<br />
Now, let<br />
R<br />
K1<br />
[ H ]<br />
� �<br />
s<br />
⎛<br />
ˆ<br />
[ COO ] ⎞<br />
⎜<br />
⎜˜<br />
⎝⎜<br />
[ COOH]<br />
⎠⎟<br />
then, the fraction of carboxyl groups ionised, X COO, will be:<br />
X<br />
COO ˆ<br />
ˆ<br />
COO<br />
COOH COO R<br />
[ ]<br />
�<br />
�<br />
ˆ<br />
[ ] � [ ] 1�R<br />
(2.40)<br />
(2.41)<br />
The number of surface charges generated on the BSA surface owing to<br />
the ionisation of the carboxyl groups can then be found via:<br />
Z ˆ � X ˆ � Total number of carboxyl groups on BSA surface (2.42)<br />
COO COO<br />
Similar calculations can be performed for the other amino acids <strong>and</strong><br />
the total charge number due to the acid–base equilibria on the BSA surface<br />
can then be found.<br />
ZAB � ∑ �� � ∑ ��<br />
(2.43)<br />
To solve these equations, the pK a values of the amino acid groups in<br />
their environment at the BSA surface need to be known. The pK a values<br />
for the amino acid groups on BSA are available in the literature [51, 52].<br />
If this data were not available for the specific protein being investigated,<br />
general data is available in the literature for the intrinsic pK a values of<br />
the ionisable amino acid groups found in proteins [52]. These pK a values<br />
can be substantially different from the pK a values of the free amino acids.<br />
The above equations have shown how the complex acid–base equilibria<br />
of the amino acid surface groups may be described, but the average
net molecular charge of the BSA molecule will also depend on whether<br />
some surface groups on the molecule will also be involved with ionic<br />
equilibria with other ions in the electrolyte solution. For BSA, chloride<br />
binding will occur [53]. This chloride binding may be described by [54]:<br />
Z<br />
Cl<br />
� �<br />
( )<br />
( )<br />
� �zeψ<br />
440 � [ Cl ] exp o<br />
kT<br />
� �zeψ<br />
1� 44 � [ Cl ] exp o<br />
kT<br />
2.3 INTERACTION FORCES 51<br />
�<br />
�zeψo<br />
( kT)<br />
�zeψo<br />
( kT<br />
33 γ[<br />
Cl ] exp<br />
�<br />
�<br />
1�1. 1 γ [ Cl ] exp )<br />
(2.44)<br />
where � is the activity coefficient of the chloride ion at the particle surface<br />
(determined from activity coefficient data for NaCl solutions).<br />
Therefore, the overall surface charge number of a BSA molecule is:<br />
Z T �Z AB �Z �<br />
Cl<br />
(2.45)<br />
When solving the PBE using charge regulation as a boundary condition,<br />
the compact part of the double layer around the BSA molecule needs<br />
to be taken into account. Figure 2.4(b) illustrates the model assumed<br />
for the compact part of the double layer. This model can be termed the<br />
Zeroth-order Stern model [55], as we have a zone of thickness, d, that is<br />
devoid of ions <strong>and</strong> represents a distance of closest approach to the particle<br />
surface of charge density � o. From electroneutrality,<br />
�o �� �d<br />
From the solution of the PBE, � d can be determined as:<br />
� ε ε � d<br />
d � o r<br />
dr<br />
r�a� d<br />
(2.46)<br />
(2.47)<br />
The surface potential of the particle, � o, may also be determined by<br />
allowing for the capacitance of the fluid in the compact layer around the<br />
sphere. For two concentric spheres, the capacitance, C, may be determined<br />
using [56]:<br />
aα<br />
C � 4πεoεr α � a<br />
(2.48)<br />
where a is the radius of the inner sphere <strong>and</strong> � (� a � d) is the radius of<br />
the outer sphere.<br />
The capacitance can also be evaluated from the surface charge density as:<br />
C Q<br />
2<br />
4πa<br />
�o<br />
� �<br />
∆V<br />
� � �<br />
o d<br />
(2.49)
52 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
Combining equations (2.48) <strong>and</strong> (2.49),<br />
�<br />
o<br />
oad<br />
�<br />
ε ε a d<br />
�<br />
� �<br />
�<br />
( )<br />
o r<br />
d<br />
(2.50)<br />
The surface potential can therefore be established if a, d, �d, �o, �εo <strong>and</strong> εr are known. The value for the dielectric constant for the compact layer, εr, is unknown, but for the model used here, it has been shown that the bulk<br />
dielectric constant for water may be used if d is in the range 0.1–0.3 nm<br />
[55]. Thus, from an initial guess for �d, �d (<strong>and</strong> thus �o) may be deter-<br />
mined from the solution of the non-linear PBE (which gives d dr r a d<br />
�/ � � ).<br />
From these values, the surface potential, � o, can be evaluated from equation<br />
(2.50) <strong>and</strong> then the number of charges on the BSA molecule may be<br />
determined from the charge regulation model. Using Z T, a value for the<br />
surface charge density may be recalculated.<br />
�<br />
o<br />
ZT e<br />
�<br />
2<br />
4πa<br />
(2.51)<br />
The two values for the surface charge calculated, via the PBE <strong>and</strong> the<br />
surface charge model, may then be compared for a given � d <strong>and</strong> an iteration<br />
may be performed on � d until the surface charge calculated via both<br />
methods is equivalent. In this way, for given dispersion conditions (pH,<br />
ionic strength, concentration), the value of the potential at the OHP may<br />
be determined. This potential is widely considered to be the zeta potential<br />
of the molecule [57]. The calculated value of the zeta potential may<br />
be compared to the value obtained experimentally from electrophoresis<br />
measurements on dilute BSA dispersions [58].<br />
Once the PBE has been solved, calculation of the interaction energy<br />
may be performed using an appropriate model.<br />
In the case of concentrated colloidal dispersions, however, the interaction<br />
energy between particles (as in a gel layer) is multiparticle in nature,<br />
so modification of the two-body interaction has to be made in order to<br />
allow for multiparticle interactions. A method by which the multiparticle<br />
nature of such interactions can be taken into account is to use a cell<br />
model [59] combined with a numerical solution of the non-linear PBE<br />
in spherical coordinates [60–64]. This cell model is based on the Wigner<br />
<strong>and</strong> Seitz cell model [65] that approximated the free electron energy of a<br />
crystal lattice by calculating the energy of a single crystal, since it has the<br />
same symmetry as the lattice.<br />
The concentrated colloidal dispersion can now be considered as being<br />
divided into spherical cells so that each cell contains a single particle <strong>and</strong><br />
a concentric spherical shell of an electrolyte solution, having an outer<br />
radius of certain magnitude such that the particle cell volume ratio in<br />
the unit cell is equal to the particle volume fraction throughout the entire
suspension, <strong>and</strong> the overall charge density within the cell is zero (electroneutral).<br />
This kind of approach gives a mean field approximation that<br />
accounts for multiparticle interactions to yield the configurational electrostatic<br />
free energy per particle [78]. By equating the configurational free<br />
energy with the pairwise summation of forces in hexagonal arrays, an<br />
expression for the repulsive force between two particles can be obtained,<br />
which implicitly takes into account the multiparticle effect [60].<br />
1 ⎛ ⎛<br />
o ze�β ( D)<br />
⎞ ⎞<br />
FR ( D) � Sβ ( D) n kT ⎜<br />
⎜cosh<br />
⎜ � 1<br />
3 ⎝⎜<br />
⎝⎜<br />
kT ⎠⎟<br />
⎠⎟<br />
(2.52)<br />
where S�(D) is the surface area of the spherical cell around the particle,<br />
no is the ion number concentration, z is the valence of the ions, e is the<br />
elementary electronic charge <strong>and</strong> ��(D) is the potential at the surface of<br />
the spherical cell.<br />
In order to evaluate the above equation, the size of the cell <strong>and</strong> the<br />
potential at the cell surface need to be known. The radius of the fluid<br />
shell can be determined with the volume fraction approach [62]. The<br />
potential at the outer boundary of the cell may be determined by solving<br />
the non-linear PBE in spherical coordinates numerically, using the electroneutrality<br />
boundary condition at the cell surface (i.e. d�/ dr r�β<br />
� 0)<br />
<strong>and</strong> the appropriate boundary condition at the particle surface.<br />
The interaction energy may now be determined using:<br />
V ( D) � � F ( D) dD<br />
R<br />
∫<br />
D<br />
∞<br />
R<br />
(2.53)<br />
This interaction energy implicitly takes multiparticle interactions into<br />
account.<br />
2.3.3 dlVo theory<br />
2.3 INTERACTION FORCES 53<br />
DLVO theory is named after Derjaguin <strong>and</strong> L<strong>and</strong>au [66] <strong>and</strong> Verwey<br />
<strong>and</strong> Overbeek [24], who were responsible for its development during<br />
the 1940s. This theory describes the forces present between charged surfaces<br />
interacting through a liquid medium. It combines the effects of the<br />
London dispersion van der Waals attraction <strong>and</strong> the electrostatic repulsion<br />
due to the overlap of the double layer of counterions. The central<br />
concept of the DLVO theory is that the total interaction energy of two<br />
surfaces or particles is given by the summation of the attractive <strong>and</strong><br />
repulsive contributions. This can be written as:<br />
VT � VA � VR<br />
(2.54)
54 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
where the total interaction energy V T is expressed in terms of the repulsive<br />
double layer interaction energy V R <strong>and</strong> the attractive London–van<br />
der Waals energy V A. For a measurement made between a spherical colloid<br />
probe <strong>and</strong> a plain surface, this can be adapted to give the relationship<br />
for a normalised force:<br />
F<br />
R<br />
�2π( V � V )<br />
A R<br />
(2.55)<br />
Contrary to the double layer interaction, the van der Waals interaction<br />
energy is mostly insensitive to variations in electrolyte strength <strong>and</strong> pH.<br />
Additionally, the van der Waals attraction must always be greater than<br />
the double layer repulsion at extremely small distances since the interaction<br />
energy satisfies a power law (i.e. V A � � D �n ), whereas the double<br />
layer interaction energy remains finite or increases far more slowly<br />
within the same small separation range.<br />
DLVO theory was challenged by the existence of long-range attractive<br />
electrostatic forces between particles of like charge. The established theory<br />
of colloidal interactions predicts that an isolated pair of like-charged<br />
colloidal spheres in an electrolyte should experience a purely repulsive<br />
screened electrostatic (Coulombic) interaction. The experimental evidence,<br />
however, indicates that the effective interparticle potential can<br />
have a long-range attractive component in more concentrated suspensions<br />
[67, 68] <strong>and</strong> for particles confined by charged glass walls [69, 70].<br />
The explanations for the observation are divided <strong>and</strong> debatable. One of<br />
the arguments [71] demonstrated that the attractive interaction measured<br />
between like-charged colloidal spheres near a wall can be accounted for<br />
by a non-equilibrium hydrodynamic effect, which was proved by both<br />
analytical results <strong>and</strong> Brownian dynamics simulations.<br />
2.3.3.1 Measurement of DLVO Forces Using Atomic Force Microscopy<br />
The following section will review a number of papers describing<br />
experiments to measure van der Waals <strong>and</strong> electrostatic double layer<br />
forces between particles using the colloidal probe technique, but should<br />
not be considered as a complete coverage of published information. For<br />
further reading, there are a number of excellent reviews that may be recommended<br />
[13, 21, 72–74].<br />
The first use of AFM to measure interactions between a colloidal particle<br />
<strong>and</strong> a surface was reported by Ducker <strong>and</strong> colleagues, where a silica<br />
sphere was allowed to interact with a silica surface in aqueous NaCl<br />
solutions [4, 75]. Force measurements were fitted with theoretical force<br />
laws for DLVO theory, combining van der Waals <strong>and</strong> electrostatic double<br />
layer interactions. For separations greater than 3 nm, observations agreed<br />
favourably with the conventional DLVO theory. However, at closer
2.3 INTERACTION FORCES 55<br />
separations, deviations did occur. This was attributed by the authors to<br />
an effect of the roughness of the surfaces involved, leading to a deviation<br />
from the assumed idealised geometry.<br />
There are a number of examples of the AFM being used along with the<br />
colloidal probe technique to probe van der Waals interactions between<br />
surfaces in a variety of media. Whilst most such studies examine a combination<br />
of van der Waals along with other surface forces, there are a few<br />
that concentrate on van der Waals measurements. Milling <strong>and</strong> colleagues<br />
[76] observed the interactions between colloidal gold spheres <strong>and</strong> poly(tetrafluoroethylene)<br />
(PTFE) surfaces in a variety of different liquids. Hamaker<br />
constants were found by fitting the appropriate force laws to the experimental<br />
data. As noted by the authors, the experimentally determined<br />
values for the Hamaker constant were only in agreement qualitatively<br />
with those calculated from theory. In addition, theoretical values varied<br />
significantly, depending upon whether an amorphous or crystalline form<br />
of PTFE was considered. For the majority of liquids used, which were of<br />
a polar nature (water, ethanol <strong>and</strong> dimethyl sulfoxide), only attractive<br />
forces were measured. In the case of dimethyl formamide, which was also<br />
a polar molecule, only repulsive interactions were observed. For these<br />
interactions, only water had been expected to show an attractive interaction.<br />
The authors concluded that in the case of the other polar liquids<br />
studied that showed attraction, the surfaces must have become contaminated<br />
by charged groups, leading to interactions that were not purely of<br />
a van der Waals nature. For interactions taking place in the perfluoroalkanes<br />
perfluorohexane <strong>and</strong> perfluorocyclohexane, as well as in air, shortrange<br />
interactions were also attractive. This was as predicted from theory,<br />
but the authors were unable to obtain experimentally determined values<br />
for A H. In the other low-polarity solvents examined (cyclohexane, dodecane,<br />
p-xylene <strong>and</strong> bromobenzene), all interactions were found to be repulsive,<br />
which was qualitatively in accordance with theoretical A H values calculated<br />
assuming amorphous PTFE surfaces.<br />
Karamen et al. [77] studied the interactions between model aluminium<br />
oxide surfaces in 1 mM solutions of potassium chloride as a function of<br />
pH. At separation distances greater than 5 nm, interaction forces were<br />
described very well by the traditional DLVO theory. From calculations<br />
of the surface potential from forces measured using DLVO theory, it was<br />
possible to calculate an isoelectric point at approximately pH 7, which<br />
was in agreement with values obtained from the literature.<br />
Many colloidal systems consist of particles with self-assembled monolayers<br />
covalently bound to the surface. Such a system is inherently more<br />
complex than those having interactions between single materials, as if the<br />
outer layer is thin, then interactions between the underlying surfaces will<br />
occur as well as between the monolayers. Kane <strong>and</strong> Mulvaney [78] studied<br />
the interactions between gold spheres <strong>and</strong> surfaces before <strong>and</strong> after
56 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
the deposition of mercaptoundecanoic acid (MUA). Force measurements<br />
at different electrolyte concentrations <strong>and</strong> pH values were fitted with a<br />
modified DLVO theory to determine surface potentials <strong>and</strong> Hamaker<br />
constants. The adsorption of the MUA itself was found to substantially<br />
decrease the magnitude of the dispersive forces between the gold surfaces.<br />
It was also found that the degree of ionisation of the surface carboxy<br />
groups was very low. The authors applied two models to explain<br />
these results. It was concluded that either a thicker-than-usually-expected<br />
Stern layer was present or that strong competitive binding of cations was<br />
the cause. Surface interactions between particles <strong>and</strong> surfaces with more<br />
complex attached monolayers, such as proteins, have also been investigated.<br />
<strong>Bowen</strong> et al. [79] investigated systems of silica microspheres <strong>and</strong><br />
surfaces co-incubated to give an adsorbed layer of bovine serum albumin<br />
(BSA) at different values of pH <strong>and</strong> NaCl concentrations. Figure 2.5<br />
shows a plot of representative force curves obtained as a BSA-coated<br />
silica sphere approached a BSA-coated silica surface in aqueous NaCl<br />
solutions of differing ionic strengths. As the ionic strength was increased,<br />
both the range <strong>and</strong> magnitude of repulsive interactions increased. For<br />
the three highest NaCl concentrations, experimental observations were<br />
in good agreement with the values predicted from DLVO theory <strong>and</strong> zeta<br />
potential values. The zeta potentials for the BSA were calculated using a<br />
site-dissociation-site-binding approach from the amino acid sequence [58,<br />
80]. At the lowest ionic strength examined (0.0002 M NaCl), the observed<br />
values were higher than those predicted from theory.<br />
Normalised force, F/R (mN/m)<br />
1<br />
0.1<br />
0.01<br />
0.0001<br />
0 5 10 15 20 25 30<br />
Distance [nm]<br />
FIGuRE 2.5 Plot of normalised force versus separation distance plot for silica surfaces<br />
coated with monolayers of BSA at four concentrations of NaCl at pH 8.0: (�) 0.1 M; (�) 0.01 M;<br />
(�) 0.001 M; (�) 0.0002 M. The lines are theoretical predictions from DLVO theory.
Interaction between colloidal particles <strong>and</strong> filtration media is of much<br />
interest industrially, due to the propensity of dispersed particles to foul<br />
membrane surfaces <strong>and</strong> reduce their efficiency [81, 82]. As under conditions<br />
in which membrane filtration most often occurs both the membrane<br />
<strong>and</strong> dispersed colloids <strong>and</strong> other particles carry surface charges,<br />
the strength <strong>and</strong> sign of electrical double layer interactions are of great<br />
importance. The forces involved in the approach of a colloidal particle to<br />
a membrane surface are essentially a balance between electrical double<br />
layer interactions <strong>and</strong> hydrodynamic forces. Studies have been made of<br />
double layer interactions using model silica colloidal spheres <strong>and</strong> polymeric<br />
microfiltration membranes [83]. Solutions of NaCl were prepared<br />
of varying ionic strengths at a single pH value (pH 8.0). Representative<br />
force curves for each approach made by the silica probe to membrane surfaces<br />
are shown in Figure 2.6. At all ionic strengths, forces were significantly<br />
repulsive at all separation distances, showing that the electrostatic<br />
double layer repulsion is dominant in all cases. The range that the repulsive<br />
interactions first become detectable increased from approximately<br />
10 nm at 10 �1 M NaCl to approximately 35 nm at 10 �4 M. When measurements<br />
were made in high purity water, this increased to 60 nm. The<br />
change in the magnitude of the forces with decreasing ionic strength is<br />
in accordance with electrical double layer theory, with the decay lengths<br />
of the measured force curves comparable to the Debye charge screening<br />
lengths of the different ionic strength solutions under study. The decrease<br />
F/R [mN/m]<br />
20<br />
15<br />
10<br />
5<br />
2.3 INTERACTION FORCES 57<br />
10�3 10<br />
M NaCl<br />
�2 10<br />
M NaCl<br />
�1 M NaCl<br />
10 �4 M NaCl<br />
0<br />
0 5 10 15 20 25 30 35 40 45<br />
Distance [nm]<br />
FIGuRE 2.6 Plot showing normalised force versus separation distance for the approach<br />
of a silica colloid probe towards a microfiltration membrane surface at different salt concentrations.<br />
As the concentration of NaCl is increased, repulsive forces on approach decrease<br />
owing to charge screening effects.
58 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
in repulsive interaction between the particles <strong>and</strong> the membrane with the<br />
increase in the salt concentration has implications for membrane filtration<br />
processes, but especially for the desalination treatment of water. Brant <strong>and</strong><br />
Childress [84] studied the long-range interaction forces between colloids<br />
of a variety of materials (silica, alumina <strong>and</strong> polystyrene) with reverse<br />
osmosis membranes <strong>and</strong> developed an extended DLVO theory that added<br />
the effect of acid–base interactions, with AFM experiments used to validate<br />
the extended theory. It was found that for interactions between two<br />
hydrophilic surfaces, the extended theory fitted the data better than classical<br />
theory, demonstrating the importance of acid–base interactions for<br />
this configuration. For interactions between hydrophobic materials, the<br />
extended theory was not significantly different from the classical theory,<br />
as would be expected for interactions between non-polar materials.<br />
2.3.4 Solvation Forces<br />
The DLVO theory successfully explains the long-range interaction<br />
forces observed in a large number of systems (colloids, surfactant solutions,<br />
lipid bilayers etc.) in terms of the electrical double layer <strong>and</strong><br />
London–van der Waals forces. However, when two surfaces or particles<br />
approach closer than a few nanometres, the interactions between two solid<br />
surfaces in a liquid medium fail to be accounted for by DLVO theory. This<br />
is because the theories of van der Waals <strong>and</strong> double layer forces discussed<br />
in the previous sections are both continuum theories, described on the<br />
basis of the bulk properties of the intervening solvent such as its refractive<br />
index, dielectric constant <strong>and</strong> density, whereas the individual nature<br />
of the molecules involved, such as their discrete size, shape <strong>and</strong> chemistry,<br />
was not taken into consideration by DLVO theory. Another explanation<br />
for this is that other non-DLVO forces come into play, although the<br />
physical origin of such forces is still somewhat obscure [85, 86]. These<br />
additional forces can be monotonically repulsive, monotonically attractive<br />
or even oscillatory in some cases. And these forces can be much stronger<br />
than either of the two DLVO forces at small separations [2, 87].<br />
To underst<strong>and</strong> how the additional forces arise between two surfaces a<br />
few nanometers apart, we need to start with the simplest but most general<br />
case of inert spherical molecules between two smooth surfaces, first<br />
considering the way solvent molecules order themselves at a solid–liquid<br />
interface, then considering how this structure corresponds to the presence<br />
of a neighbouring surface <strong>and</strong> how this in turn brings about the shortrange<br />
interaction between two surfaces in the liquid. Usually, the liquid<br />
structure close to an interface is different from that in the bulk. For many<br />
liquids, the density profile normal to a solid surface oscillates around<br />
the bulk density, with a periodicity of a molecular diameter in a narrow<br />
region near the interface. This region typically extends over several
2.3 INTERACTION FORCES 59<br />
molecular diameters. Within this range, the molecules are ordered in layers<br />
according to some theoretical work <strong>and</strong> particularly computer simulations<br />
[88, 89] as well as experimental observations [90, 91]. When two<br />
such surfaces approach each other, one layer of molecules after another is<br />
squeezed out of the closing gap. The geometric constraining effect of the<br />
approaching wall on these molecules <strong>and</strong> attractive interactions between<br />
the surface <strong>and</strong> liquid molecules hence create the solvation force between<br />
the two surfaces. For simple spherical molecules between two hard,<br />
smooth surfaces, the solvation force is usually a decaying oscillatory<br />
function of distance. For molecules with asymmetric shapes or whose<br />
interaction potentials are anisotropic or not pairwise additive, the resulting<br />
solvation force may also have a monotonically repulsive or attractive<br />
component. When the solvent is water, they are referred to as hydration<br />
forces. Solvation forces depend both on the chemical <strong>and</strong> physical properties<br />
of the surfaces being considered, such as the wettability, crystal<br />
structure, surface morphology <strong>and</strong> rigidity <strong>and</strong> on the properties of the<br />
intervening medium.<br />
The hydration force is one of the most widely studied <strong>and</strong> controversial<br />
non-DLVO forces, a strong short-range force that decays exponentially<br />
with the distance, D, between the surfaces [92, 93]:<br />
F ( D) � Ke<br />
SOL<br />
� D/ l<br />
(2.56)<br />
where K � 0 relates to the hydrophilic repulsion forces, K � 0 relates to<br />
the hydrophobic attraction forces <strong>and</strong> l is the correlation length of the orientational<br />
ordering of water molecules.<br />
The concept of the hydration force emerged to explain measurements<br />
of forces between neutral lipid bilayer membranes [93]. Its presence in<br />
charged systems is controversial, but there is experimental evidence of<br />
non-DLVO forces following equation (2.56) in systems as diverse as<br />
dihexadecyldimethyl ammonium acetate surfactant bilayers [94], DNA<br />
polyelectrolyte solutions [95] <strong>and</strong> charged polysaccharides [96]. In these<br />
experiments, the hydration forces show little sensitivity to ionic strength.<br />
Many theoretical studies <strong>and</strong> computer simulations of various confined<br />
liquids, including water, have invariably led to a solvation force<br />
described by an exponentially decaying cos-function of the form<br />
[97–100]:<br />
⎛ πD⎞<br />
FSOL ( D) � f0cos<br />
⎜ e<br />
⎝⎜<br />
� ⎠⎟<br />
2 �D/<br />
D0<br />
m<br />
(2.57)<br />
where F SOL is the force per unit area; f 0 is the force extrapolated to separation<br />
distance, D � 0; � m is the molecular diameter <strong>and</strong> D 0 is the characteristic<br />
decay length.
60 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
A repulsive force dominant at short ranges between silica surfaces in<br />
aqueous solutions of NaCl has been reported by Grabbe <strong>and</strong> Horn [101],<br />
which was also found to be independent of electrolyte concentration over<br />
the range investigated. They attributed this force to a hydration repulsion<br />
resulting from hydrogen bonding of water to the silica surface, <strong>and</strong> fitted<br />
the additional component to a sum of two exponentials to work out the<br />
formula for the hydration forces in the system.<br />
The physical mechanisms underlying the hydration force are still a<br />
matter for debate. One possible mechanism is the anomalous polarisation<br />
of water near the interfaces, which completely alters its dielectric<br />
response [102–104]. These theories imply an electrostatic origin of the<br />
hydration force. However, other authors report [105] that there is no<br />
evidence for a significant structuring of water layers near interfaces, or<br />
a perturbation of its dielectric response, as envisaged by previous theories.<br />
Instead, they suggest that the repulsive forces are due to entropic<br />
(osmotic) repulsion of thermally excited molecular groups that protrude<br />
from the surfaces [106]. This theory explains many experimental observations<br />
in neutral systems [107], but its validity in charged systems is not<br />
certain. Given the available evidence from experiments <strong>and</strong> simulations,<br />
it is not possible to reach a definitive conclusion on the precise role of<br />
these mechanisms in determining the hydration forces. Until recently,<br />
computer simulations of water films coated with ionic surfactants<br />
showed that protrusions are not significant in these systems [108]. On<br />
the other h<strong>and</strong>, computer simulations show that water has an anomalous<br />
dielectric behaviour near charged interfaces [109], but the observed electrostatic<br />
fields obviously differ from the predictions of electrostatic theories<br />
on hydration forces [103, 110]. The effect of this anomalous dielectric<br />
behaviour of water on the electrostatic force between surfaces or interfaces<br />
is still unknown.<br />
The use of AFM to measure solvation forces between surfaces has been<br />
of some interest amongst researchers in recent years for the purposes of<br />
both studying the phenomena between closely interacting probes <strong>and</strong><br />
surfaces that may cause artefacts when imaging with an AFM <strong>and</strong> investigating<br />
the effect of these solvation forces on the interactions between<br />
colloidal particles. O’Shea <strong>and</strong> colleagues [111] observed an oscillatory<br />
force on close approach of a sharp AFM probe to a graphite surface in<br />
water, octamethylcyclotetra siloxane (OMCTS), <strong>and</strong> dodecanol. A series<br />
of repulsive barriers were observed as approach was made for OMCTS<br />
<strong>and</strong> dodecanol, with each repulsive barrier becoming successively larger,<br />
owing to the closer-bound solvation layers becoming harder to displace.<br />
The authors calculated that the energy required to remove the solvation<br />
molecules was 5–25 times kT for OMCTS <strong>and</strong> 5–1000 times kT for<br />
dodecanol, suggesting that only a small number of molecules were being<br />
displaced. In addition, it was noted that the distance between successive
2.3 INTERACTION FORCES 61<br />
oscillatory peaks compared well with the sizes of the molecules of interest,<br />
suggesting that each solvation shell consisted of a single layer of<br />
molecules. A later set of measurements in OMCTS <strong>and</strong> dodecanol on<br />
HOPG reached a similar conclusion [112, 113]. No displacement of solvation<br />
layers could be observed in water owing to the relatively large range<br />
<strong>and</strong> magnitude of the attractive jump-in event observed. In a later set of<br />
experiments [114], an oscillating lever was used to probe the solvation<br />
forces in OMCTS <strong>and</strong> dodecanol, allowing the mechanical compliance of<br />
the solvation shells to be measured. Here, periodic changes in the amplitude<br />
of lever oscillations were observed, which was explained as being<br />
because of an increase in the effective viscosity of the solutions when the<br />
surfaces were in close approach. It was noted that the presence of oscillatory<br />
structural forces even at the highly curved geometries present, even<br />
when using a sharp AFM probe, could have a detrimental influence on<br />
the very high resolution imaging of surfaces when using AFM.<br />
Solvation forces were measured in primary alcohols between silicon<br />
nitride probes <strong>and</strong> mica <strong>and</strong> HOPG surfaces by Franz <strong>and</strong> Butt [115]. For<br />
the measurements made against the hydrophilic mica, no solvation oscillation<br />
forces were observed at separation distances of less than 4 nm. At<br />
distances less than this, repulsive maxima followed by sudden jump-in<br />
events were observed. These repeated maxima were ascribed to the presence<br />
of solvation shells, with at least two of these layers present. These<br />
solvation forces were determined to be greater in magnitude than the<br />
attractive van der Waals forces present. It was noted that the period of<br />
the force oscillations increased linearly with chain length, with the period<br />
greater than the chain length. It was concluded that for the measurements<br />
made on mica, the molecules did not take on a flat conformation,<br />
but were at least partially upright. On the hydrophobic HOPG surface,<br />
measurements in 1-propanol <strong>and</strong> 1-pentanol oscillations had a period<br />
of 0.45 nm, suggesting that on the hydrophobic surface they did lie flat<br />
against the surface.<br />
Valle-Delgado <strong>and</strong> colleagues [116] investigated the solvation forces<br />
in water between hydrophilic silica spheres versus plane silica surfaces.<br />
A repulsive force was measured (�2 nm) at ranges shorter than the<br />
observed double layer repulsion <strong>and</strong> attractive van der Waals forces.<br />
A number of theoretical models were applied to the experimental data<br />
<strong>and</strong> it was concluded that for the situation under investigation, the observations<br />
were best explained by the formation <strong>and</strong> rupture of hydrogen<br />
bonds between SiOH groups on the silica surfaces <strong>and</strong> a single layer of<br />
water molecules.<br />
Jarvis et al. [117] used a carbon nanotube to probe solvation forces in<br />
water against a self-assembled monolayer bound to a gold surface, which<br />
had been characterised by imaging using the nanotube probe. When the<br />
forces were normalised, they were found to be in reasonable agreement
62 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
with previous measurements undertaken with the surface force apparatus<br />
[118]. The authors concluded that the forces scaled along with the surface<br />
interaction dimensions between the mesoscale <strong>and</strong> nanoscale [117].<br />
2.3.5 Steric Interaction Forces<br />
For molecules attached to a solid surface in a liquid environment,<br />
chains with a degree of freedom to move will tend to dangle out into the<br />
solution where they remain thermally mobile. On approach of two polymer-covered<br />
surfaces, the entropy of confining these dangling chains<br />
results in a repulsive entropic force, which, for overlapping polymer<br />
molecules, is known as the steric or overlap repulsion. In ancient Egypt,<br />
people already knew how to keep ink stabilised by dispersing soot particles<br />
in water, incubated with gum arabicum or egg-adsorbed polymers,<br />
which, in the first case, is a mixture of polysaccharide <strong>and</strong> glycoprotein<br />
<strong>and</strong> in the second mainly the protein albumin, which works through this<br />
steric repulsion. However, steric repulsion does not necessarily have to<br />
be due to polymeric molecules; layers of small molecules can have the<br />
same effect, albeit at a much shorter range.<br />
Steric stabilisation of dispersions is very important in many industrial<br />
processes. This is because colloidal particles that normally coagulate in<br />
a solvent can often be stabilised by adding a small amount of polymer<br />
to the dispersing medium. Such polymer additives are known as protectives<br />
against coagulation <strong>and</strong> they lead to the steric stabilisation of a colloid.<br />
Both synthetic polymers <strong>and</strong> biopolymers (e.g. protein, gelatine) are<br />
widely used in both non-polar <strong>and</strong> polar solvents (e.g. in paints, toners,<br />
emulsions, cosmetics, pharmaceuticals, processed food, soils, lubricants).<br />
Theories of steric interactions are not well developed. There is no simple,<br />
comprehensive theory available as steric forces are complicated <strong>and</strong><br />
difficult to describe [119–121]. The magnitude of the force between surfaces<br />
coated with polymers depends on the quantity or coverage of polymer on<br />
each surface, on whether the polymer is simply adsorbed from solution (a<br />
reversible process) or irreversibly grafted onto the surfaces <strong>and</strong> finally on<br />
the quality of the solvent [119, 122]. Different components contribute to the<br />
force, <strong>and</strong> which component dominates the total force is situation specific.<br />
For interactions in poor <strong>and</strong> theta solvents, there are some theories<br />
available for low <strong>and</strong> high surface coverage. In the case of the low coverage<br />
where there is no overlap or entanglement of neighbouring chains,<br />
the repulsive energy per unit area is a complex series <strong>and</strong> roughly exponential<br />
[123–126]. As for the high coverage of end-grafted chains, the<br />
thickness of the brush layer increases linearly with the length of the polymer<br />
molecule. Once two brush-bearing surfaces are close enough to each<br />
other, there is a repulsive pressure between them, <strong>and</strong> this force can be<br />
approximated by the Alex<strong>and</strong>er–de Gennes theory [119, 127, 128].
Studies of forces between model alumina surfaces in electrolyte solutions<br />
have noted the existence of steric interactions at basic (�8) pH,<br />
most probably due to the formation of a hydrated gel layer at the solid/<br />
liquid interface [77, 129] providing an extra resistance to hard contact<br />
being made. At low pH values, the surface forces were described by classical<br />
DLVO theory, but at high pH with the presence of the hydrated gel<br />
layer, deviations from DLVO theory occurred at close separations. Polat<br />
et al. [129] measured frictional (lateral) forces in addition to forces normal<br />
to the surface. Differences in the normal forces at different pH values had<br />
a significant effect on the frictional forces. At pH values where the additional<br />
repulsive force due to hydration was present, the frictional forces<br />
were significantly decreased. The authors concluded that such behaviour<br />
was likely to have implications for the rheology stability <strong>and</strong> forming<br />
behaviour of powders made from this material <strong>and</strong> potentially for some<br />
other metal oxide materials.<br />
The adsorption of electrolyte <strong>and</strong> surfactant molecules from solution<br />
to solid surfaces is also likely to result in the addition of steric forces on<br />
close approach between surfaces. Studies of the surface forces between<br />
gold-coated silica spheres mounted on AFM cantilevers <strong>and</strong> plane gold<br />
surfaces in aqueous solutions of gold chloride, sodium chloride <strong>and</strong> trisodium<br />
citrate found that when the citrate <strong>and</strong> chloride anions were added<br />
together a short-range steric barrier appeared [130]. The range of this steric<br />
barrier was equal to the size of two citrate anions, suggesting that the<br />
citrate was coating each opposing surface with a monolayer. Meagher <strong>and</strong><br />
colleagues studied the interactions between silica microspheres <strong>and</strong> �-alumina<br />
plane surfaces in the presence of different combinations of electrolyte,<br />
polyelectrolyte <strong>and</strong> surfactant [131]. In the presence of aqueous electrolyte<br />
alone, long-range forces compared well with DLVO theory at all separations.<br />
At high pH values, all forces measured were repulsive. As the pH<br />
was decreased, the repulsion was reduced, eventually starting to show features<br />
of attraction as pH was reduced to below the isoelectric point for the<br />
alumina (pH 5.5). When the polyelectrolyte sodium poly(styrene sulfonate)<br />
was added, with <strong>and</strong> without the presence of the cationic surfactant cetyltrimethylammonium<br />
bromate (CTAB) (at concentrations below its critical<br />
micelle concentration), at pH values equal to or higher than the isoelectric<br />
point of alumina, an additional repulsive force was observed at approaches<br />
of 3 nm or less. This deviation from the DLVO theory was observed as a<br />
short-range repulsive force that varied in its magnitude <strong>and</strong> range.<br />
2.3.6 Hydrophobic Interaction Forces<br />
2.3 INTERACTION FORCES 63<br />
A hydrophobic surface usually has no polar or ionic groups or hydrogen-bonding<br />
sites so that there is no affinity for water <strong>and</strong> the surface<br />
to bond together. Ordinary water in bulk is significantly structured
64 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
because of hydrogen bonding between the water molecules. The cooperative<br />
nature of this bonding [132] means that quite large clusters of<br />
hydrogen-bonded water molecules can form although they may continually<br />
form <strong>and</strong> break down in response to thermal energy fluctuations.<br />
The orientation of water molecules in contact with a hydrophobic molecule<br />
is entropically unfavourable; therefore, two such molecules tend<br />
to come together simply by attracting each other. As a result, the entropically<br />
unfavoured water molecules are expelled into the bulk <strong>and</strong> the<br />
total free energy of the system is reduced accordingly. The presence of<br />
a hydrophobic surface could restrict the natural structuring tendency<br />
of water by imposing a barrier that prevents the growth of clusters in a<br />
given direction. Similar effects occur between two hydrophobic surfaces<br />
in water. Water molecules confined in a gap between two such surfaces<br />
would thus be unable to form clusters larger than a certain size. For an<br />
extremely narrow gap, this could be a serious limitation <strong>and</strong> result in an<br />
increased free energy of the water in comparison with that in bulk. In<br />
other words, this would give rise to an attractive force between hydrophobic<br />
surfaces as a consequence of water molecules migrating from the<br />
gap to the bulk water where there are unrestricted hydrogen-bonding<br />
opportunities <strong>and</strong> a lower free energy.<br />
Attraction between hydrophobic surfaces has been measured directly<br />
[133] <strong>and</strong> can be of surprisingly long range, up to about 80 nm [134]. The<br />
attraction was much stronger than the van der Waals force <strong>and</strong> of much<br />
greater range. The interaction of filaments of hydrophobised silica was<br />
measured by Rabinovich <strong>and</strong> Derjaguin [135]. They found an attractive<br />
force at large separation distances, one to two orders of magnitude<br />
greater than van der Waals attraction. There have been many experimental<br />
measurements of the interaction force between various hydrophobic<br />
surfaces in aqueous solutions. These studies have found that the hydrophobic<br />
force between two macroscopic surfaces is of extraordinarily<br />
long range, decaying exponentially with a characteristic decay length of<br />
1–2 nm in the range 0–10 nm, <strong>and</strong> then more gradually further out, <strong>and</strong><br />
this force can be much stronger than those predicted on the basis of van<br />
der Waals interaction, especially between hydrocarbon surfaces for which<br />
the Hamaker constant is quite small.<br />
It is now well established that a long-range (�10 nm) attractive force<br />
operates between hydrophobic surfaces immersed in water <strong>and</strong> aqueous<br />
solutions [136]. Unfortunately, so far, no generally accepted theory has<br />
been developed for these forces, but the hydrophobic force is thought to<br />
arise from overlapping solvation zones as two hydrophobic species come<br />
together [2]. In fact, Eriksson et al. [137] have used a square-gradient variational<br />
approach to show that the mean field theory of repulsive hydration<br />
forces can be modified to account for some aspects of hydrophobic<br />
attraction. Conversely, Ruckenstein <strong>and</strong> Churaev suggest a completely
2.3 INTERACTION FORCES 65<br />
different origin that attributes the attraction to the coalescence of vacuum<br />
gaps at the hydrophobic surfaces [138].<br />
In recent years, a number of studies have been undertaken to study<br />
the effect of nanoscale bubbles found on hydrophobic surfaces to<br />
explain the hydrophobic forces reported in many surface force measurements.<br />
These bubbles may be present on the hydrophobic surfaces<br />
when first immersed in aqueous solutions, or may form from dissolved<br />
gases after immersion. Considine et al. [139], when undertaking measurements<br />
between two latex spheres, noted a large attractive force with<br />
a range much in excess of that expected from van der Waals attraction<br />
<strong>and</strong> independent of electrolyte concentration. They observed that degassing<br />
of solutions reduced the range of this attractive force significantly.<br />
Re-gassing of the solution restored the original range of this force, showing<br />
the influence of dissolved gas on the measurement of hydrophobic<br />
forces. Mahnke <strong>and</strong> colleagues [140] studied both the effect of dissolved<br />
gas <strong>and</strong> the degree of hydrophobicity of surfaces on the range of attractive<br />
forces. When surfaces were used that gave large (�25 nm) jump-in<br />
events, degassing of the intervening medium reduced the range of measurable<br />
attractive forces. For combinations of surfaces that gave short<br />
jump-in events, degassing of the solutions had little effect. It can be concluded<br />
that the long-range hydrophobic attraction can be accounted for<br />
by the presence of nanobubbles attached to the hydrophobic surfaces.<br />
The presence of such bubbles has been confirmed by the use of tapping<br />
mode AFM on hydrophobised silicon wafers [141, 142]. Bubble-like features<br />
observed on these surfaces show a high phase contrast with the<br />
rest of the surface, suggesting very different mechanical properties to<br />
the silicon surface. Force curves taken at the sites of these putative nanobubbles<br />
show much greater attractive <strong>and</strong> adhesive forces with a hydrophobic<br />
AFM probe than when taken from a bare area of the surface.<br />
Zhang <strong>and</strong> colleagues examined the effects of degassing <strong>and</strong> liquid temperature<br />
on the number <strong>and</strong> density on the surface of the nanobubbles<br />
[143]. Degassing of water <strong>and</strong> ethanol under vacuum reduced the surface<br />
density of nanobubbles to a very significant extent. Increasing the<br />
temperature of the fluid also increased the number <strong>and</strong> size of the nanobubbles<br />
on the surface, a phenomena witnessed by another group who<br />
also observed the spontaneous appearance of nanobubbles as the liquid<br />
temperature was increased [144]. For measurements where the hydrophobic<br />
force being measured is a result of the interaction of bubbles on<br />
the opposing surfaces, the forces measured are actually capillary forces<br />
rather than true hydrophobic forces, albeit capillary forces present as a<br />
result of the hydrophobicity of the surfaces. For those who wish to read<br />
further into the origins of forces measured between hydrophobic surfaces<br />
in aqueous solutions, the authors would like to draw attention to a number<br />
of reviews in the literature [145–147].
66 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
2.3.7 Effect of Hydrodynamic drag on AFM Force<br />
Measurements<br />
When performing measurements with particles in liquid, the effects of<br />
hydrodynamic drag on both the particle <strong>and</strong> also the cantilever may be significant,<br />
depending on the velocities at which measurements are taken. In<br />
particular, the effects of confinement of the liquid between two particles or<br />
between two surfaces are important. As the distance between the two surfaces<br />
decreases, the finite drainage time for the confined liquid produces a<br />
force dependent on the separation distance, the velocity of the approach as<br />
well as the viscosity <strong>and</strong> density of the fluid medium. Ignoring any contribution<br />
of drag on the cantilever to this effect, the hydrodynamic force, F H,<br />
for a sphere approaching a plane surface is given by a modified version of<br />
Stoke’s law for the drag force on a sphere [148, 149]:<br />
F � 6πν�r �<br />
H s c<br />
(2.58)<br />
where υ is the velocity of the probe, � is the dynamic viscosity of the<br />
surrounding fluid, r s is the radius of the spherical particle <strong>and</strong> � c is a correction<br />
applied to Stoke’s law to account for the presence of the confining<br />
wall in close proximity.<br />
where<br />
� c<br />
4<br />
� � sinh�<br />
3<br />
∞<br />
∑<br />
c c<br />
n�1<br />
n( n � 1)<br />
( 2n � 1)( 2n � 3)<br />
⎡ 2sinh( 2n � 1) �c<br />
+ ( 2n + 1)<br />
sinh2�<br />
⎤<br />
⎢<br />
c<br />
− 1⎥<br />
⎢ 2 2<br />
⎣⎢<br />
4sinh ( n � 0. 5) �c � ( 2n � 1) sinh2�<br />
⎥<br />
c ⎦⎥<br />
�1<br />
D � rs<br />
D r<br />
� cosh �ln<br />
rs<br />
rs<br />
⎛<br />
⎧<br />
⎞<br />
⎪<br />
⎪⎛<br />
⎞<br />
⎜ ⎨<br />
⎪ ⎜<br />
⎝⎜<br />
⎠⎟<br />
⎪<br />
⎪⎝⎜<br />
⎠⎟<br />
⎩⎪<br />
� s<br />
�<br />
⎡<br />
⎢⎛D<br />
� r ⎞<br />
⎜ s<br />
⎢⎜<br />
⎢⎜⎝<br />
rs<br />
⎠⎟<br />
⎣<br />
2<br />
⎤⎫⎪<br />
⎥<br />
⎪<br />
� 1⎥<br />
⎬<br />
⎪<br />
⎥⎪<br />
⎦⎪<br />
⎭⎪<br />
(2.59)<br />
(2.60)<br />
where D is the distance between the plane surface <strong>and</strong> closest point of the<br />
sphere. For measurements where r � �h, then this can be simplified to:<br />
F<br />
H<br />
rs<br />
�<br />
D<br />
6πν�<br />
2<br />
(2.61)<br />
In most cases, contributions due to interactions between the surface<br />
<strong>and</strong> the cantilever itself can be neglected. However, depending on the<br />
size of the colloid probe <strong>and</strong> the speed at which the experiment is carried<br />
out, this may not always be the case. Vinogradova et al. [150] studied the
2.4 AdHESION FORCES MEASUREd By AFM 67<br />
effect of sphere size <strong>and</strong> probe speed on the contribution of drag on the<br />
AFM cantilever to measured hydrodynamic forces. They found that for<br />
the cantilever used that at speeds of 7.5 �m s �1 <strong>and</strong> for spheres of radii<br />
less than 3 �m, the cantilever did measurably contribute to the observed<br />
effects. Obviously, for faster speeds, the minimum sphere size to prevent<br />
cantilever contributions confounding sphere–plane measurements will<br />
increase. For most operating conditions, the authors recommended using<br />
colloid probes of no less than 5 �m in radius.<br />
2.4 AdHESIon FoRCES MEASuREd by AFM<br />
Adhesive forces measured as the pull-off force in AFM measurements<br />
represent the sum of all the interaction forces occurring in the contact<br />
regime. This includes long-range forces such as van der Waals <strong>and</strong> electrostatic<br />
forces; capillary forces (if measurements are undertaken in air);<br />
solvation forces; hydrophobic interactions <strong>and</strong> steric interactions as well<br />
as any chemical bonding between groups present on the surfaces.<br />
2.4.1 Contact Mechanics <strong>and</strong> Adhesion<br />
The magnitude of the adhesion between two surfaces is dependent<br />
upon the contact area at the junction between the surfaces as well as<br />
the interaction forces themselves. The contact area will depend upon<br />
the mechanical deformation of the materials due to the applied force<br />
<strong>and</strong> material properties. The underst<strong>and</strong>ing of the relationship between<br />
adhesive forces <strong>and</strong> contact mechanics as generally applied in the field of<br />
force microscopy is dependent upon the works of Johnson, Kendal <strong>and</strong><br />
Roberts (JKR theory) [151] <strong>and</strong> Derjaguin, Muller <strong>and</strong> Toporov (DMT<br />
theory) [152] some decades ago. According to the JKR model of contact<br />
mechanics, for a sphere–plane system, the adhesion (pull-off) force can<br />
be related to the work of adhesion, contact area <strong>and</strong> mechanical compliance<br />
of the interacting surfaces by:<br />
JKR 3<br />
ac<br />
2<br />
FAd � πRpWa<br />
�<br />
2<br />
R 3<br />
2<br />
p<br />
6πW<br />
E*<br />
a<br />
(2.62)<br />
where R p is the radius of the probe sphere, W a is the work of adhesion per<br />
unit area <strong>and</strong> a c is the contact radius. The parameter a 2 c/R is the sample<br />
deformation <strong>and</strong> �. E* is the reduced Young’s modulus for the system:<br />
2<br />
1 3 1 1<br />
s<br />
E* 4 Es E f<br />
�<br />
⎛<br />
⎜ � ν � ν<br />
⎜ �<br />
⎜<br />
⎝⎜<br />
2<br />
f<br />
⎞<br />
⎠⎟<br />
(2.63)
68 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
where E s, E f <strong>and</strong> v s, v f are the Young’s modulus <strong>and</strong> Poisson ratios for the<br />
sphere <strong>and</strong> flat surface respectively. The JKR model applies well for large<br />
probes with soft samples <strong>and</strong> large adhesions. For the case of relatively<br />
small tips with surfaces with high Young’s moduli <strong>and</strong> low adhesion, the<br />
DMT model may apply better. For the DMT theory, a slightly different<br />
relationship is found:<br />
a F R W<br />
FAd R W<br />
R R E<br />
DMT<br />
c<br />
p a<br />
� p a �<br />
p<br />
�<br />
2 ( 2π<br />
)<br />
2π<br />
3 2<br />
*<br />
p<br />
2<br />
3<br />
(2.64)<br />
The JKR <strong>and</strong> DMT models are really descriptions of the two ends of a<br />
continuum. Tabor suggested a way of deciding which of these two models<br />
would be the best to apply to a certain situation. From the following<br />
relationship, if the factor � R is greater than unity, then the JKR theory<br />
would be best applied, with DMT being appropriate for � R values less<br />
than unity [153, 154]:<br />
� R<br />
2<br />
a p<br />
0 3<br />
1<br />
⎞ 3<br />
* ⎟<br />
⎛<br />
⎜W<br />
R<br />
� 2. 92 ⎜<br />
⎝⎜<br />
E z ⎠<br />
(2.65)<br />
where � R is effectively the ratio of the elastic deformation due to the applied<br />
load <strong>and</strong> adhesion to the effective range of the surface forces (z 0) [153, 155].<br />
For intermediate systems where values of � R are close to unity, then a treatment<br />
using the Maugis–Dugdale theory is more appropriate [155].<br />
The AFM has been used to measure adhesive forces between particles<br />
<strong>and</strong> process surfaces. One important example is adhesion between particles,<br />
including both inorganic colloidal particles <strong>and</strong> bacterial cells, <strong>and</strong><br />
filtration membranes. This interaction is of great importance when considering<br />
the fouling <strong>and</strong> biofouling of such surfaces. Particles adhere to<br />
the process membranes <strong>and</strong> reduce flow through the membrane, greatly<br />
reducing filtration, the efficiency <strong>and</strong> working lifetime of the membranes.<br />
The process testing of new membranes is potentially expensive <strong>and</strong> time<br />
consuming. The quantification of adhesion forces between colloids <strong>and</strong><br />
membranes can provide an important contribution to developing the<br />
theoretical prediction <strong>and</strong> optimisation <strong>and</strong> control of many engineering<br />
separation processes. As a result, the development of AFM methods to<br />
quantify the adhesion of different materials to membranes of different<br />
compositions can potentially be very useful for membrane manufacturers<br />
<strong>and</strong> engineers [156]. When particle sizes are greater than the pore size<br />
in the absence of repulsive double layer interactions, such particles may<br />
plug the pores very effectively, leading to a catastrophic loss in filtration<br />
flux. Of the many established membrane characterisation techniques,
2.4 AdHESION FORCES MEASUREd By AFM 69<br />
only the colloid probe method can measure the adhesive forces between<br />
particles <strong>and</strong> membrane surfaces <strong>and</strong> hence allow prediction of the membrane<br />
fouling properties of the particles. In addition, the ability to make<br />
measurements in liquid allows the matching of observation conditions to<br />
those that occur in practice.<br />
Studies have also been carried out on the adhesion forces between<br />
calcium carbonate crystals <strong>and</strong> stainless steel surfaces. The adhesion of<br />
CaCO 3 to steel process equipment surfaces plays an important role in the<br />
formation of scale deposits in desalination plant equipment, as well as in<br />
domestic appliances such as kettles <strong>and</strong> other household appliances. Al-<br />
Anezi et al. [157] measured the adhesion between CaCO 3 probes <strong>and</strong> stainless<br />
steel surfaces of different grades of roughness. In Figure 2.7 is an SEM<br />
image showing an example of a CaCO 3 crystal mounted on an AFM cantilever.<br />
The roughest surface had the lowest adhesion than the two smoother<br />
surfaces examined. The effect of liquid environment was also examined.<br />
Adhesive forces measured in sea water to which 2 ppm of an anti-scaling<br />
agent had been added were significantly more reduced than when measured<br />
in sea water alone. In addition, the proportion of measured force<br />
curves that showed no observable adhesion rose from 1% of all obtained<br />
force curves in sea water to 57% with the addition of the anti-scale agent.<br />
FIGuRE 2.7 SEM image of a CaCO 3 crystal mounted upon an AFM cantilever for use in<br />
determining adhesion with stainless steel surfaces.
70 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
2.5 EFFECt oF RouGHnESS on MEASuREd AdHESIon<br />
And SuRFACE FoRCES<br />
The continuum theories outlined above, such as the JKR <strong>and</strong> DMT<br />
theories, assume that perfectly smooth surfaces come into contact [2].<br />
Unfortunately, the degree of roughness of particles <strong>and</strong> surfaces is quite<br />
variable, <strong>and</strong> often, except for when studying molecularly smooth surfaces,<br />
some account may be needed to be taken of roughness. The presence<br />
of asperities on surfaces coming into contact serves, in most cases,<br />
to effectively decrease the contact area. There are a number of models<br />
that have been developed to account for the effect of surface roughness<br />
on measured adhesion when trying to infer various properties of interactions<br />
from adhesion forces. These generally use some measure of roughness,<br />
such as the root mean square (rms) roughness of the surface, or<br />
values for mean asperity size to account for the reduced contact areas<br />
due to the presence of surface asperities [158–165]. As well as serving<br />
to keep the two surfaces separated, the angle at which asperities on the<br />
particle surfaces approach each other may also affect the effective contact<br />
area <strong>and</strong> thus the measured adhesion [166]. Rabinovich et al. [162] determined<br />
that surface roughness values as small as 1.6 nm rms were significant<br />
<strong>and</strong> could reduce adhesion values by as much as fivefold from that<br />
expected.<br />
In addition, rough particles make determination of the radius of probe<br />
particles, <strong>and</strong> hence normalisation by particle size, difficult. Larson <strong>and</strong><br />
colleagues [167] determined an effective probe radius when performing<br />
measurements between TiO 2 colloids <strong>and</strong> crystal surfaces by fitting force<br />
data to an electrical double layer model, with surface potential values<br />
determined independently. This allowed calculation of an effective probe<br />
radius that was used to normalise all subsequent force measurements.<br />
AbbREVIAtIonS And SyMbolS<br />
a Effective hard sphere particle radius in solution m<br />
ac Contact radius m<br />
ai Particle radius of molecule i M<br />
a VW van der Waals gas constant N m 4 mol �2<br />
A131 Effective Hamaker constant in medium 3 J<br />
AH Hamaker constant (in vacuum or in medium) J<br />
bVW van der Waals gas constant m3 mol�1 c Velocity of light in a vacuum (2.998 � 108 ) m s�1
C Capacitance C V �1<br />
C D Debye interaction constant J m 6<br />
C K Keesom interaction constant J m 6<br />
C L London interaction constant J m 6<br />
C T Interaction constant J m 6<br />
C UV Cauchy Plot parameter �<br />
d Distance to OHP (surface of shear) m<br />
D Surface to surface separation distance m<br />
D jtc Jump-in distance m<br />
D 0 Characteristic decay length m<br />
e Elementary charge (1.602 � 10 �19 ) C<br />
E* Reduced Young’s modulus Pa s �2<br />
E s Young’s modulus for the sphere Pa<br />
E f Young’s modulus for the flat surface Pa<br />
f 0 Force extrapolated to separation distance, D � 0 N<br />
F Force between surfaces/particles N<br />
JKR<br />
FAd DMT<br />
FAd ABBREvIATIONS ANd SyMBOLS 71<br />
Adhesion pull-off force (JKR model) N<br />
Adhesion pull-off force (DMT model) N<br />
FH Hydrodynamic force for a sphere approaching a<br />
plane surface<br />
N<br />
FR Repulsive interparticle force N<br />
FSOL Hydration force N<br />
h Planck’s constant (6.626 � 10 �34 ) m 2 kg s �1<br />
¯h Planck’s constant divided by 2� (6.626 �<br />
10�34 /2�)<br />
J s rad�1 k Boltzmann constant (1.380 � 10�23 ) J K�1 kc Spring constant of the cantilever N m�1 keff Effective spring constant of the cantileversample<br />
system<br />
N m�1 ks Stiffness of the sample N m�1 K Constant in force equation N<br />
Ki Equilibrium constant for ionisation reaction i M<br />
l Correlation length of the orientational ordering<br />
of water molecules<br />
m<br />
n Number of gas molecules mol
72 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
n o Ion number concentration in bulk m �3<br />
n oi<br />
Refractive index of component i �<br />
p Pressure N m �2<br />
Q Charge on plates C<br />
r Interparticle separation (centre to centre) m<br />
r s Radius of the spherical particle m<br />
R Gas constant J mol �1 K �1<br />
R i Radius of sphere i m<br />
R p Radius of the probe sphere m<br />
R t Radius of curvature of the probe tip m<br />
S � Surface area of spherical shell m<br />
T Absolute temperature K<br />
u i Dipole moment of molecule or atom i C m<br />
v Velocity of the probe m s �1<br />
v f Poisson ratios for a flat surface �<br />
v i Orbiting frequency of electron i s �1<br />
v s Poisson ratio for a sphere �<br />
V Volume of container m 3<br />
V A Attractive interaction energy J<br />
V R Repulsive interaction energy J<br />
V T Total interaction energy J<br />
�V Voltage difference V<br />
w (r) Interaction potential J<br />
W(D) Interaction energy per unit area as a function of<br />
distance<br />
J m�2 Wa Work of adhesion between two bodies J<br />
xjtc Cantilever deflection due to the jump-in m<br />
X ˆ<br />
COO Fraction of carboxyl groups ionised �<br />
zi Valence �<br />
z0 Effective range of the surface forces m<br />
Z� Charge number due to negative groups on the<br />
BSA surface<br />
�<br />
Z� Charge number due to positive groups on the<br />
BSA surface<br />
�<br />
ZAB Charge number from acid–base equilibria �
GREEk SyMBOLS 73<br />
Z cl � Number of bound chloride ions �<br />
Z COO ˆ Charge number due to ionisation of carboxyl<br />
groups<br />
Z T Total charge number of BSA molecule �<br />
GREEk SyMbolS<br />
� Hydrodynamic radius (� a � d) m<br />
� 0i Electronic polarisability of molecule i m 3<br />
� Activity coefficient of chloride ion in NaCl solution �<br />
� i Reduced surface potential i �<br />
� ik Interfacial energy between components i <strong>and</strong> k J m �2<br />
ε 0 Permittivity of vacuum (8.854 � 10) C V �1 m �1<br />
ε ri Dielectric constant of component i �<br />
� Zeta potential V<br />
�i Number of atoms per unit volume of the interacting<br />
material<br />
m�3 � Distance parameter in Lennard-Jones equation m<br />
�d Charge density of diffuse double layer C m�2 �m Molecular diameter m<br />
�o Surface charge density C m�2 � Debye–Hückel parameter m�1 � Characteristic wavelength m<br />
�c Correction applied to Stoke’s law �<br />
� Viscosity of fluid N s m�2 �R Ratio of the elastic deformation to the applied load <strong>and</strong><br />
adhesion to the effective range of the surface forces<br />
� Electrostatic potential V<br />
�� Electrostatic potential at outer cell boundary V<br />
�d Electrostatic potential at OHP (� zeta potential) V<br />
�o Electrostatic potential at particle surface V<br />
wi Characteristic frequency of electromagnetic radiation rad s�1 wUV UV characteristic frequency of electromagnetic<br />
radiation<br />
rad s�1 ε Depth of the potential energy well J<br />
�
74 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
References<br />
[1] B.V. Derjaguin, Analysis of friction <strong>and</strong> adhesion IV. The theory of the adhesion of<br />
small particles, Kolloid-Zeitschrift 69 (2) (1934) 155–164.<br />
[2] J.N. Israelachvili, Intermolecular <strong>and</strong> surface forces, Academic Press, London, 1992.<br />
[3] I.U. Vakarelski, K. Higashitani, Single-nanoparticle-terminated tips for scanning probe<br />
microscopy, Langmuir 22 (2006) 2931–2934.<br />
[4] W.A. Ducker, T.J. Senden, Measurement of forces in liquids using a force microscope,<br />
Langmuir 8 (7) (1992) 1831–1836.<br />
[5] J.E. Lennard-Jones, Cohesion, Proc. Phys. Soc. 43 (5) (1931) 461–482.<br />
[6] F. London, Some characteristics <strong>and</strong> uses of molecular force, Zeitschrift Fur<br />
Physikalische Chemie-Abteilung B-Chemie Der Elementarprozesse Aufbau Der<br />
Materie 11 (2/3) (1930) 222–251.<br />
[7] F. London, The general theory of molecular forces, Trans. Faraday Soc. 33 (1937) 8–26.<br />
[8] H.C. Hamaker, The London–van der Waals attraction between spherical particles,<br />
Physica 4 (10) (1937) 1058–1072.<br />
[9] J.H. de Boer, The influence of van der Waals’ forces <strong>and</strong> primary bonds on binding<br />
energy, strength <strong>and</strong> orientation, with special reference to some artificial resins, Trans.<br />
Faraday Soc. 32 (1936) 10–37.<br />
[10] V.A. Parsegian, Van Der Waals Forces: A H<strong>and</strong>book for Biologists, Chemists,<br />
Engineering <strong>and</strong> Physicists, Cambridge University Press, 2006.<br />
[11] E.M. Lifshitz, The theory of molecular attractive forces between solids, Sov. Phys. 2 (1)<br />
(1956) 73–83.<br />
[12] P.F. Luckham, Manipulating forces between surfaces: applications in colloid science<br />
<strong>and</strong> biophysics, Adv. Colloid Interface Sci. 111 (2004) 29–47.<br />
[13] B. Capella, G. Dietler, Force–distance curves by atomic force microscopy, Surf. Sci.<br />
Rep. 34 (1999) 1–104.<br />
[14] H.B.G. Casimir, D. Polder, The influence of retardation on the London–van der Waals<br />
forces, Phys. Rev. 73 (4) (1948) 360–372.<br />
[15] B.W. Ninham, V.A. Parsegian, Van der Waals forces: special characteristics in lipid–<br />
water systems <strong>and</strong> general method of calculation based on the Lifshitz theory,<br />
Biophys. J. 10 (1970) 646–663.<br />
[16] D.B. Hough, L.R. White, The calculation of Hamaker constants from Lifshitz theory<br />
with applications to wetting, Adv. Colloid Interface Sci. 40 (1980) 8–41.<br />
[17] R.G. Horn, J.N. Israelachvili, Measurement of structural forces between two particles<br />
in a non-polar liquid, J. Chem. Phys. 75 (1981) 1400–1411.<br />
[18] W.B. Russel, D.A. Saville, W.R. Schowalter, Colloidal Dispersions, Cambridge<br />
University Press, Cambridge, 1989.<br />
[19] J. Manhanty, B.W. Ninham, Dispersion Forces (1976) Academic Press, London.<br />
[20] P.D. Ashby, L. Chen, L.C.M., Probing intermolecular forces <strong>and</strong> potentials with magnetic<br />
feedback chemical force microscopy, J. Am. Chem. Soc. 122 (2000) 9467–9472.<br />
[21] H.-J. Butt, B. Capella, M. Kapple, Force measurements with the atomic force microscope:<br />
technique, interpretation <strong>and</strong> applications, Sur. Sci. Rep. 59 (2005) 1–152.<br />
[22] S. Das, P.A. Sreeram, A.K. Raychaudhuri, A method to quantitatively evaluate the<br />
Hamaker constant using the jump-into-contact effect in atomic force microscopy,<br />
Nanotechnology 18 (2007) 035501.<br />
[23] S. Assemi, J. Nalaskowski, W.P. Johnson, Direct force measurements between carboxylate-modified<br />
latex microspheres <strong>and</strong> glass using atomic force microscopy, Colloids<br />
Surf. A Physicochem. Eng. Asp. 286 (2006) 70–77.<br />
[24] E.J.W. Verwey, J.T.G. Overbeek, Theory of the Stability of Lyophobic Colloids, Elsevier<br />
Publishing Company, Inc, Amsterdam, 1948.<br />
[25] O. Stern, The theory of the electrolytic double shift, Z. Elektrochem. 30 (1924) 508–516.<br />
[26] R.J. Hunter, Foundations of Colloid Science, second ed., Oxford University Press,<br />
Oxford, 2001.
REFERENCES 75<br />
[27] G.M. Bell, S. Levine, L.N. McCartney, Approximate methods of determining the double-<br />
layer free energy of interaction between two charged colloidal spheres, J. Colloid<br />
Interface Sci. 33 (3) (1970) 335–359.<br />
[28] J. Gregory, Interaction of unequal double layers at constant charge, J. Colloid Interface<br />
Sci. 51 (1975) 44–51.<br />
[29] R.I. Hogg, T.W. Healey, D.W. Fuerstenau, Mutual coagulation of colloidal dispersions,<br />
Trans. Faraday Soc. 62 (1966) 1638–1651.<br />
[30] G.R. Wiese, T.W. Healy, Effect of particle size on colloid stability, Trans. Faraday Soc.<br />
66 (1970) 490–500.<br />
[31] H. Ohshima, T. Kondo, Comparison of three models on double layer interaction,<br />
J. Colloid Interface Sci. 126 (1988) 382–383.<br />
[32] G. Kar, S. Ch<strong>and</strong>er, T.S. Mika, The potential energy of interaction between dissimilar<br />
electrical double layers, J. Colloid Interface Sci. 44 (1973) 347–355.<br />
[33] L.N. McCartney, S. Levine, An improvement on Derjaguin’s expression at small<br />
potentials for the double-layer interaction energy of two spherical colloidal particles,<br />
J. Colloid Interface Sci. 30 (1969) 345–354.<br />
[34] H. Muller, Zur Theorie der elektrischen Ladung und der Koagulation der Kolloide,<br />
Kolloidchemische Beihefte 26 (1928) 257.<br />
[35] N.E. Hoskin, Solution to the Poisson–Boltzmann equation for the potential distribution in<br />
the double layer of a single spherical colloidal particle, Trans. Faraday Soc. 49 (1953) 1471.<br />
[36] A.L. Loeb, P.H. Weirsema, J.T.G. Overbeek, The electrical double layer around a spherical<br />
colloid particle, MIT Press, Cambridge Mass, 1961.<br />
[37] W.R. <strong>Bowen</strong>, F. Jenner, Dynamic ultrafiltration model for charged colloidal dispersions:<br />
a Wigner–Seitz cell approach, Chem. Eng. Sci. 50 (1995) 1707.<br />
[38] W.R. <strong>Bowen</strong>, A.O. Sharif, Adaptive finite-element solution of the nonlinear Poisson–<br />
Boltzmann equation: a charged spherical particle at various distances from a charged<br />
cylindrical pore in a charged planar surface, J. Colloid Interface Sci. 187 (1997) 363.<br />
[39] J.J. Gray, B. Chiang, R.T. Bonnecaze, Colloidal particles: origin of anomalous multibody<br />
interactions, Nature 402 (1999) 705.<br />
[40] N.E. Hoskin, The interaction of 2 identical spherical colloidal particles. 1. Potential<br />
distribution, Philos. Trans. R. Soc. (Lond.) A 248 (1956) 433.<br />
[41] J.E. Ledbetter, T.L. Croxton, D.A. McQuarrie, The interaction of two charged spheres<br />
in the Poisson–Boltzmann equation, Can. J. Chem. 59 (1981) 1860.<br />
[42] C.W. Outhwaite, P. Molero, Effective interaction between colloidal particles using a<br />
symmetric Poisson–Boltzmann theory, Chem. Phys. Lett. 197 (1992) 643–648.<br />
[43] J.E. Sanchez-Sanchez, M. Lozada-Cassou, Exact numerical solution to the integral<br />
equation version of the Poisson–Boltzmann equation of two interacting spherical colloidal<br />
particles. Chem. Phys. Lett. (190) (1992) 202–208.<br />
[44] M. Strauss, T.A. Ring, H.K. <strong>Bowen</strong>, Osmotic pressure for concentrated suspensions of<br />
polydisperse particles with thick double layers, J. Colloid Interface Sci. 118 (1987) 326–334.<br />
[45] W.R. <strong>Bowen</strong>, P.M. Williams, Finite difference solution of the 2-dimensional Poisson–<br />
Boltzmann equation for spheres in confined geometries, Colloids Surf. A Physicochem.<br />
Eng. Asp. 204 (2002) 103–115.<br />
[46] M. Ospeck, S. Fraden solving the Poisson–Boltzmann equation to obtain information<br />
energies between confined, like-charged cylinders, J. Chem. Phys. 109 (1998) 9166.<br />
[47] W.R. <strong>Bowen</strong>, P.M. Williams, Dynamic ultafiltration model for proteins: a colloidal<br />
interaction approach, Biotechnol. Bioeng. 50 (2) (1996) 125–135.<br />
[48] B.W. Ninham, V.A. Parsegian, Electrostatic potential between surfaces bearing ionizable<br />
groups in ionic equilibrium with physiological saline solution, J. Theor. Biol. 31<br />
(1971) 405–428.<br />
[49] J.R. Brown, P. Shockley. In: P. Jost, H. Griffiths (Eds.), Lipid–Protein Interactions, New<br />
York, Wiley, 1982, p. 25.<br />
[50] C. Tanford, S.A. Swanson, W.S. Shore, Hydrogen ion equilibria of bovine serum albumin,<br />
J. Am. Chem. Soc. 77 (1955) 6414.
76 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
[51] R.C. Cantor, P.R. Schimmel, Biophysical Chemistry. Part II – Techniques for the Study<br />
of Biological Structure <strong>and</strong> Function, Freeman <strong>and</strong> Company, New York, 1980.<br />
[52] G. Scatchard, E.S. Black, The effect of salts on the isoionic <strong>and</strong> isoelectric points of<br />
proteins, J. Phys. Colloid Chem. 53 (1949) 88.<br />
[53] W.R. <strong>Bowen</strong>, P.M. Williams, The osmotic pressure of electrostatically stabilized colloidal<br />
dispersions, J. Colloid Interface Sci. 184 (1996) 241.<br />
[54] T.W. Healy, L.E. White, Ionizable surface group models of aqueous interfaces, Adv.<br />
Colloid Interface Sci. 9 (1978) 303.<br />
[55] R.A. Serway, Physics for Scientists <strong>and</strong> Engineers with Modern Physics, 3rd ed,<br />
Updated Version, Chicago, Saunders College Publishing, 1992, p. 714.<br />
[56] R.J. Hunter, Zeta Potential in Colloid Science, Academic Press, London, 1981.<br />
[57] S. Kuwabara, The forces experienced by r<strong>and</strong>omly distributed parallel circular cylinders<br />
or spheres in a viscous flow at small Reynolds number, J. Physical Soc. Japan 14<br />
(4) (1959) 527–532.<br />
[58] E.S. Reiner, C.J. Radke, Electrostatic interactions in colloidal suspensions – tests of<br />
pairwise additivity, AIChE J. 37 (6) (1991) 805–824.<br />
[59] N.E. Hoskin, S. Levine, The interaction of two identical spherical colloidal particles: II.<br />
The free energy, Philos. Trans. R. Soc. (Lond.) A 248 (951) (1956) 449–466.<br />
[60] L. Guldbr<strong>and</strong>, L.G. Nilsson, L.A. Nordenskiöld, Monte Carlo simulation study of electrostatic<br />
forces between hexagonally packed DNA double helices, J. Chem. Phys. 85<br />
(1986) 6687–6698.<br />
[61] E. Wigner, F. Seitz, On the constitution of metallic sodium, Phys. Rev. 43 (10) (1933)<br />
804–810.<br />
[62] B.V. Derjaguin, L.D. L<strong>and</strong>au, Theory of the stability of strongly charged lyophobic<br />
sols <strong>and</strong> of the adhesion of strongly charged particles in solutions of electrolytes, Acta<br />
Physicochim. URSS 14 (1941) 633.<br />
[63] A.E. Larsen, D.G. Grier, Like-charge attractions in metastable colloidal crystallites,<br />
Nature 385 (1997) 230–233.<br />
[64] N. Ise, T. Okubo, M. Sugimura, K. Ito, H.J. Nolte, Ordered structure in dilute solutions<br />
of highly charge polymers lattices as studied by microscopy. I. Interparticle distance<br />
as a function of latex concentration, J. Chem. Phys. 78 (1983) 536–540.<br />
[65] J.C. Crocker, D.G. Grier, Microscopic measurements of pair interaction potential of<br />
charge stabilised colloid, Phys. Rev. Lett. 73 (1994) 352–355.<br />
[66] D.G. Grier, Optical tweezers in colloid <strong>and</strong> interface science, Curr. Opin. Colloid<br />
Interface Sci. 2 (1997) 264–270.<br />
[67] T.M. Squires, M.P. Brenner, Like-charge attraction <strong>and</strong> hydrodynamic interaction,<br />
Phys. Rev. Lett. 85 (23) (2000) 4976–4979.<br />
[68] H.-J. Butt, R. Berger, E. Bonnacurso, Y. Chen, J. Wang, Impact of atomic force microscopy<br />
on interface <strong>and</strong> colloid science, Adv. Colloid Interface Sci. (2007).<br />
[69] J. Ralston, I. Larson, M.W. Rutl<strong>and</strong>, A.A. Feiler, M. Kleijn, Atomic force microscopy<br />
<strong>and</strong> direct surface force measurements (IUPAC technical report), Pure Appl. Chem. 77<br />
(2005) 2149–2170.<br />
[70] D.J. Johnson, N.J. Miles, N. <strong>Hilal</strong>, Quantification of particle–bubble interactions using<br />
atomic force microscopy: a review, Adv. Colloid Interface Sci. 127 (2) (2006) 67–81.<br />
[71] W.A. Ducker, T.J. Senden, R.M. Pashley, Direct measurement of colloidal forces using<br />
an atomic force microscope, Nature 353 (1991) 239–241.<br />
[72] A. Milling, P. Mulvaney, I. Larson, Direct measurement of repulsive van der Waals<br />
interactions using an atomic force microscope, J. Colloid Interface Sci. 180 (1996)<br />
460–465.<br />
[73] M.E. Karaman, R.M. Pashley, T.D. Waite, S.J. Hatch, H. Bustamante, A comparison of<br />
the interaction forces between model alumina surfaces <strong>and</strong> their colloidal properties,<br />
Colloids Surf. A Physicochem. Eng. Asp. 129–130 (1997) 239–255.
REFERENCES 77<br />
[74] V. Kane, P. Mulvaney, Double-layer interactions between self-assembled monolayers<br />
of �-undecanoic acid on gold surfaces, Langmuir 14 (1998) 3303–3311.<br />
[75] W.R. <strong>Bowen</strong>, N. <strong>Hilal</strong>, R.W. Lovitt, C.J. Wright, Direct measurement of interactions<br />
between adsorbed protein layers using an atomic force microscope, J. Colloid<br />
Interface Sci. 197 (1998) 348.<br />
[76] N. <strong>Hilal</strong>, H. Al-Zoubi, N.A. Darwish, A.W. Mohammed, M. Abu Arabi, A comprehensive<br />
review of nanofiltration membranes: treatment, pretreatment, modelling <strong>and</strong><br />
atomic force microscopy, Desalination 170 (2004) 281–308.<br />
[77] N. <strong>Hilal</strong>, O.O. Ogunbiyi, N.J. Miles, R. Nigmatullin, Methods employed for control<br />
of fouling in MF <strong>and</strong> NF membranes: a comprehensive review, Separ. Sci. Technol. 40<br />
(10) (2005) 1957–2005.<br />
[78] N. <strong>Hilal</strong>, W.R. <strong>Bowen</strong>, Atomic force microscopy study of the rejection of colloids by<br />
membrane pores, Desalination 150 (2002) 289–295.<br />
[79] J.A. Brant, A.E. Childress, Membrane–colloid interactions: comparison of extended DLVO<br />
predictions with AFM force measurements, Environ. Eng. Sci. 19 (6) (2002) 413–427.<br />
[80] V. Bergeron, Forces <strong>and</strong> structure in thin liquid soap films, J. Phys. Condens. Matter 11<br />
(1999) R215–R238.<br />
[81] B.W. Ninham, On progress in forces since the DLVO theory, Adv. Colloid Interface Sci.<br />
83 (1999) 1–17.<br />
[82] H.K. Christenson, Non-DLVO forces between surfaces – solvation, hydration <strong>and</strong> capillary<br />
effects, J. Dispers. Sci. Technol. 9 (2) (1988) 171–206.<br />
[83] D.Y.C. Chan, D.J. Mitchell, B.W. Ninham, A. Pailthorpe, Short-range interactions<br />
mediated by a solvent with surface adhesion, Mol. Phys. 35 (6) (1978) 1669–1679.<br />
[84] I.K. Snook, D. Henderson, Monte–Carlo study of a hard-sphere fluid near a hard wall,<br />
J. Chem. Phys. 68 (5) (1978) 2134–2139.<br />
[85] A.K. Doerr, M. Tolan, T. Seydel, W. Press, The interface structure of thin liquid hexane<br />
films, Physica B Condens. Matter 248 (1998) 263–268.<br />
[86] C.J. Yu, A.G. Richter, J. Kmetko, A. Datta, P. Dutta, X-ray diffraction evidence of ordering<br />
in a normal liquid near the solid–liquid interface, Europhys. Lett. 50 (4) (2000) 487–493.<br />
[87] N.V. Churaev, B.V. Derjaguin, Inclusion of structural forces in the theory of stability of<br />
colloids <strong>and</strong> films, J. Colloid Interface Sci. 103 (2) (1985) 542–553.<br />
[88] S. Leikin, V.A. Parsegian, D.C. Rau, R.P. R<strong>and</strong>, Hydration forces, Annu. Rev. Phys.<br />
Chem. 44 (1993) 369–395.<br />
[89] V.A. Parsegian, R.P. R<strong>and</strong>, N.L. Fuller, Direct osmotic stress measurements of hydration<br />
<strong>and</strong> electrostatic double-layer forces between bilayers of double-chained ammonium<br />
acetate surfactants, J. Phys. Chem. 95 (1991) 4777–4782.<br />
[90] D.C. Rau, B. Lee, V.A. Parsegian, Measurement of the repulsive force between polyelectrolyte<br />
molecules in ionic solution: hydration forces between parallel DNA double<br />
helices, Proc. Natl. Acad. Sci. U.S.A. 81 (1984) 2621–2625.<br />
[91] D.C. Rau, V.A. Parsegian, Direct measurement of forces between linear polysaccharides<br />
xanthan <strong>and</strong> schizophyllan, Science 249 (1990) 1278–1281.<br />
[92] R. Evans, A.O. Parry, Liquids at interfaces – what can a theorist contribute, J. Phys.<br />
Condens. Matter 2 (1990) SA15–SA32.<br />
[93] I.K. Snook, W. van Megen, Calculation of solvation forces between solid particles<br />
immersed in a simple liquid, J. Chem. Soc. Faraday Trans. II 77 (1) (1981) 181–190.<br />
[94] P. Tarazona, L. Vicente, A model for density oscillations in liquids between solid walls,<br />
Mol. Phys. 56 (3) (1985) 557–572.<br />
[95] W. van Megan, I.K. Snook, Solvent structure <strong>and</strong> solvation forces between solid bodies,<br />
J. Chem. Soc. Faraday Trans. II 75 (1979) 1095.<br />
[96] A. Grabbe, R.G. Horn, Double-layer <strong>and</strong> hydration forces measured between silica sheets<br />
subjected to various surface treatments, J. Colloid Interface Sci. 157 (2) (1993) 375–383.<br />
[97] S. Marčelja, Hydration in electrical double layer, Nature 385 (1997) 689–690.
78 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
[98] E. Ruckenstein, M. Manciu, The coupling between the hydration <strong>and</strong> double layer<br />
interactions, Langmuir 18 (2002) 7584–7593.<br />
[99] S. Leikin, A.A. Kornyshev, Theory of hydration forces. Nonlocal electrostatic interaction<br />
of neutral surfaces, J. Chem. Phys. 92 (11) (1990) 6890–6898.<br />
[100] J. Israelachvili, H. Wennerström, Role of hydration <strong>and</strong> water structure in biological<br />
<strong>and</strong> colloidal interactions, Nature 379 (1996) 219–225.<br />
[101] D. Leckb<strong>and</strong>, J. Israelachvili, Intermolecular forces in biology, Q. Rev. Biophys. 34 (2)<br />
(2001) 105–267.<br />
[102] J.N. Israelachvili, H. Wennerström, Hydration or steric forces between amphiphilic<br />
surfaces? Langmuir 6 (1990) 873–876.<br />
[103] J. Faraudo, F. Bresme, Origin of the short-range, strong repulsive force between ionic<br />
surfactant layers, Phys. Rev. Lett. 94 (7) (2005) 077802.<br />
[104] J. Faraudo, F. Bresme, Anomalous dielectric behavior of water in ionic Newton black<br />
films, Phys. Rev. Lett. 92 (2004) 236102.<br />
[105] D.W.R. Gruen, S. Marčelja, Spatially varying polarization in water, J. Chem. Soc.<br />
Faraday Trans. 2 79 (1983) 225–242.<br />
[106] S.J. O’Shea, M.E. Well<strong>and</strong>, T. Rayment, Solvation forces near a graphite surface measured,<br />
Appl. Phys. Lett. 60 (19) (1992) 2356–2358.<br />
[107] R. Lim, S.F.Y. Li, S.J. O’Shea, Solvation forces using sample-modulation atomic force<br />
microscopy, Langmuir 18 (2002) 6116–6124.<br />
[108] L.T.W. Lim, A.T.S. Wee, S.J. O’Shea, Effect of tip size on force measurement in atomic<br />
force microscopy, Langmuir (2008).<br />
[109] S.J. O’Shea, M.E. Well<strong>and</strong>, Atomic force microscopy at solid–liquid interfaces,<br />
Langmuir 14 (1998) 4186–4197.<br />
[110] V. Franz, H.-J. Butt, Confined liquids: solvation forces in liquid alcohols between<br />
solid surfaces, J. Phys. Chem. B 106 (2002) 1703–1708.<br />
[111] J.J. Valle-Delgado, J.A. Molina-Bolivar, F. Galisteo-Gonzalez, M.J. Galvez-Ruiz, A.<br />
Feiler, M.W. Rutl<strong>and</strong>, Hydration forces between silica surfaces: experimental data<br />
<strong>and</strong> predictions from different theories, J. Chem. Phys. 123 (3) (2005) 034708.<br />
[112] S.P. Jarvis, T. Uchihashi, T. Ishida, H. Tokumoto, Y. Nakayama, Local solvation shell<br />
measurement in water using a carbon nanotube probe, J. Phys. Chem. B 104 (26)<br />
(2000) 6091–6094.<br />
[113] J. Israelachvili, R.M. Pashley, Molecular layering of water at surface <strong>and</strong> origin of<br />
repulsive hydration forces, Nature 306 (1983) 249–250.<br />
[114] P.G. de Gennes, Polymers at an interface: a simplified view, Adv. Colloid Interface<br />
Sci. 27 (1987) 189–209.<br />
[115] F.T. Hesselin, Theory of stabilization of dispersions by adsorbed macromolecules. 1.<br />
Statistics of change of some configurational properties of adsorbed macromolecules<br />
on approach of an impenetrable interface, J. Phys. Chem. 75 (1) (1971) 65.<br />
[116] A. Vrij, Polymers at interfaces <strong>and</strong> interactions in colloidal dispersions, Pure Appl.<br />
Chem. 48 (4) (1976) 471–483.<br />
[117] S.T. Milner, T.A. Witten, M.E. Cates, Theory of the grafted polymer brush, Macromolecules<br />
21 (8) (1988) 2610–2619.<br />
[118] K. Ingersent, J. Klein, P. Pincus, Interactions between surfaces with adsorbed polymers –<br />
poor solvent. 2. Calculations <strong>and</strong> comparison with experiment, Macromolecules 19 (5)<br />
(1986) 1374–1381.<br />
[119] K. Ingersent, J. Klein, P. Pincus, Forces between surfaces with adsorbed polymers –<br />
theta solvent. 3. Calculations <strong>and</strong> comparison with experiment, Macromolecules 23<br />
(2) (1990) 548–560.<br />
[120] J.N. Israelachvili, M. Tirrell, J. Klein, Y. Almog, Forces between 2 layers of adsorbed<br />
polystyrene immersed in cyclohexane below <strong>and</strong> above the theta-temperature,<br />
Macromolecules 17 (1984) 204.<br />
[121] J.W. Cahn, Critical point wetting, J. Chem. Phys. 66 (1977) 3667.
REFERENCES 79<br />
[122] P.G. de Gennes, Stabilité de films polymère/solvant, Les Comptes rendus de<br />
l’Académie des sciences (Paris) 300 (1985) 839.<br />
[123] G. Rossi, P. Pincus, Interactions between unsaturated-polymer adsorbed surfaces,<br />
Europhys. Lett. 5 (1988) 641.<br />
[124] M. Polat, K. Sato, T. Nagaoka, K. Watari, Effect of pH <strong>and</strong> hydration on the normal<br />
<strong>and</strong> lateral interaction forces between alumina surfaces, J. Colloid Interface Sci. 304<br />
(2) (2006) 378–387.<br />
[125] S. Biggs, P. Mulvaney, C.F. Zukoski, F. Grieser, Study of anion adsorption at the gold–<br />
aqueous solution interface by atomic force microscopy, J. Am. Chem. Soc. 116 (1994)<br />
9150–9157.<br />
[126] L. Meagher, G. Maurdev, M.L. Gee, Interaction forces between a bare silica surface<br />
<strong>and</strong> an alpha-alumina surface bearing adsorbed polyelectrolyte <strong>and</strong> surfactant,<br />
Langmuir 18 (7) (2002) 2649–2657.<br />
[127] M.C.R. Symons, Liquid water – the story unfolds, Chem Br. 25 (5) (1989) 491–494.<br />
[128] J.N. Israelachvili, R.M. Pashley, Measurement of the hydrophobic interaction between<br />
two hydrophobic surfaces in aqueous electrolyte solutions, J. Colloid Interface Sci. 98<br />
(1984) 500–514.<br />
[129] P.M. Claesson, H.K. Christenson, Very long range attraction between uncharged<br />
hydrocarbon <strong>and</strong> fluorocarbon surfaces in water, J. Phys. Chem. 92 (1988) 1650–1655.<br />
[130] Y.I. Rabinovich, B.V. Derjaguin, Interaction of hydrophobized filaments in aqueous<br />
electrolyte solutions, Colloids Surf. 30 (1988) 243–251.<br />
[131] J.N. Israelachvili, R.M. Pashley, The hydrophobic interaction is long range, decaying<br />
exponentially with distance, Nature 300 (1982) 341–342.<br />
[132] J.C. Eriksson, S. Ljunggren, P.M. Claesson, A phenomenological theory of long-range<br />
hydrophobic attraction forces based on a square-gradient variational approach,<br />
J. Chem. Soc. Faraday Trans. II 85 (1989) 163–176.<br />
[133] E. Ruckenstein, N. Churaev, A possible hydrodynamic origin of the forces of hydrophobic<br />
attraction, J. Colloid Interface Sci. 147 (2) (1991) 535–538.<br />
[134] R.F. Considine, R.A. Hayes, R.G. Horn, Forces measured between latex spheres in<br />
aqueous electrolyte: non-DLVO behaviour <strong>and</strong> sensitivity to dissolved gas, Langmuir<br />
15 (1999) 1657–1659.<br />
[135] J. Mahnke, J. Stearnes, R.A. Hayes, D. Fornasiero, J. Ralston, The influence of dissolved<br />
gas on the interactions between surfaces of different hydrophobicity in aqueous<br />
media, Phys. Chem. Chem. Phys. 1 (1999) 2793–2798.<br />
[136] N. Ishida, K. Higashitani, Interaction forces between chemically modified hydrophobic<br />
surfaces evaluated by AFM – the role of nanoscopic bubbles in the interactions,<br />
Min. Eng. 19 (2006) 719–725.<br />
[137] N. Ishida, T. Inoue, M. Miyahara, K. Higashitani, Nano bubbles on a hydrophobic<br />
surface in water observed by tapping-mode atomic force microscopy, Langmuir 16<br />
(2000) 6377–6380.<br />
[138] X.H. Zhang, X.D. Zhang, S.T. Lou, Z.X. Zhang, J.L. Sun, J. Hu, Degassing <strong>and</strong> temperature<br />
effects on the formation of nanobubbles at the mica/water interface,<br />
Langmuir 20 (2004) 3813–3815.<br />
[139] S. Yang, S.M. Dammer, N. Bremond, H.J.W. Z<strong>and</strong>livet, E.S. Kooij, D. Lohse, Characterization<br />
of nanobubbles on hydrophobic surfaces in water, Langmuir 23 (2007) 7072–7077.<br />
[140] P. Attard, Nanobubbles <strong>and</strong> the hydrophobic attraction, Adv. Colloid Interface Sci.<br />
104 (2003) 75–91.<br />
[141] H.K. Christenson, P.M. Claesson, Direct measurements of the force between hydrophobic<br />
surfaces in water, Adv. Colloid Interface Sci. 91 (2001) 391–436.<br />
[142] E.E. Meyer, K.J. Rosenberg, J. Israelachvili, Recent progress in underst<strong>and</strong>ing hydrophobic<br />
interactions, Proc. Natl. Acad. Sci. U.S.A. 103 (43) (2006) 15739–15746.<br />
[143] H. Brenner, The slow motion of a sphere through a fluid towards a plane surface,<br />
Chem. Eng. Sci. 16 (3–4) (1961) 242–251.
80 2. MEASUREMENT OF PARTICLE ANd SURFACE INTERACTIONS<br />
[144] V.S.J. Craig, C. Neto, In situ calibration of colloid probe cantilevers in force microscopy:<br />
hydrodynamic drag on a sphere approaching a wall, Langmuir 17 (2001) 6018–6022.<br />
[145] O.I. Vinogradova, H.-J. Butt, G.E. Yakubov, F. Feuillebois, Dynamic effects on force<br />
measurements. I. Viscous drag on the atomic force microscope cantilever, Rev. Sci.<br />
Instrum. 72 (2001) 2330–2339.<br />
[146] K.L. Johnson, K. Kendall, A.D. Roberts, Surface energy <strong>and</strong> contact of elastic solids,<br />
Proc. R. Soc. Lond. A Math. Phys. Sci. 324 (1558) (1971) 301–313.<br />
[147] B.V. Derjaguin, V.M. Muller, Y.P. Toporov, Effect of contact deformations on the adhesion<br />
of particles, J. Colloid Interface Sci. 53 (2) (1975) 314–326.<br />
[148] F.L. Leite, P.S.P. Herrmann, Application of atomic force spectroscopy (AFS) to studies<br />
of adhesion phenomena, J. Adhes. Sci. Technol. 19 (3–5) (2005) 365–405.<br />
[149] D. Tabor, Surface forces <strong>and</strong> surface interactions, J. Colloid Interface Sci. 58 (1) (1976)<br />
2–13.<br />
[150] K.L. Johnson, J.A. Greenwood, An adhesion map for the contact of elastic spheres, J.<br />
Colloid Interface Sci. 192 (1997) 326–333.<br />
[151] N. <strong>Hilal</strong>, W.R. <strong>Bowen</strong>, L. Al-Khatib, O.O. Ogunbiyi, A review of atomic force microscopy<br />
applied to cell interactions with membranes, Chem. Eng. Res. Des. 84 (A4)<br />
(2006) 282–292.<br />
[152] K. Al-Anezi, D.J. Johnson, N. <strong>Hilal</strong>, An atomic force microscope study of calcium<br />
carbonate adhesion to desalination process equipment: effect of anti-scale agent,<br />
Desalination 220 (2008) 359–370.<br />
[153] T.S. Chow, Size-dependent adhesion of nanoparticles on rough substrates, J. Phys.<br />
Condens. Matter 15 (2003) L83–L87.<br />
[154] K. Cooper, A. Gupta, S. Beaudoin, Simulation of the adhesion of particles to surfaces,<br />
J. Colloid Interface Sci. 234 (2001) 284–292.<br />
[155] K. Cooper, N. Ohler, A. Gupta, S. Beadoin, Analysis of contact interactions between a<br />
rough deformable colloid <strong>and</strong> a smooth substrate, J. Colloid Interface Sci. 222 (2000)<br />
63–74.<br />
[156] C.S. Hodges, L. Looi, J.A.S. Cleaver, M. Ghadiri, Use of the JKR model for calculating<br />
adhesion between rough surfaces, Langmuir 20 (2004) 9571–9576.<br />
[157] Y.I. Rabinovich, J.J. Adler, A. Ata, R.K. Singh, B.M. Moudgil, Adhesion between<br />
nanoscale rough surfaces. II. Measurement <strong>and</strong> comparison with theory, J. Colloid<br />
Interface Sci. 232 (2000) 17–24.<br />
[158] Y.I. Rabinovich, J.J. Adler, A. Ata, R.K. Singh, B.M. Moudgil, Adhesion between<br />
nanoscale rough surfaces. I. Role of asperity geometry, J. Colloid Interface Sci. 232<br />
(2000) 10–16.<br />
[159] J.Y. Waltz, The effect of surface heterogeneities on colloidal forces, Adv. Colloid<br />
Interface Sci. 74 (1998) 119–168.<br />
[160] R.R. Dagastine, M. Bevan, L.R. White, D.C. Prieve, Calculation of van der Waals<br />
forces with diffuse coatings: applications to roughness <strong>and</strong> adsorbed coatings, J.<br />
Adhes. 80 (2004) 365–394.<br />
[161] A. Méndez-Vilas, M.L. González-Martín, M.J. Nuevo, Some geometrical considerations<br />
about the influence of topography on the adhesion force as measured by AFM<br />
on curved surfaces, Appl. Surf. Sci. 238 (2004) 9–13.<br />
[162] I. Larson, C.J. Drummond, D.Y.C. Chan, F. Grieser, Direct force measurements<br />
between TiO 2 surfaces, J. Am. Chem. Soc. 115 (1993) 11885–11890.
C H A P T E R<br />
3<br />
Quantification of<br />
Particle–Bubble<br />
Interactions Using Atomic<br />
Force Microscopy<br />
<strong>Nidal</strong> <strong>Hilal</strong> <strong>and</strong> Daniel Johnson<br />
O U T L I N E<br />
3.1 Introduction 82<br />
3.2 Particle–Bubble Interactions 82<br />
3.3 Determination of Particle–Bubble Separation 86<br />
3.4 Determination of Contact Angle from Force–Distance Curves 88<br />
3.5 Effect of Surface Preparation on Particle–Bubble Interactions 91<br />
3.5.1 Effect of Particle Surface Chemistry on Particle–Bubble Interactions 91<br />
3.5.2 Effect of Surfactant on Particle–Bubble Interactions 94<br />
3.6 Effect of Loading Force on Particle–Bubble Interactions 97<br />
3.7 Effect of Hydrodynamics on Particle–Bubble Interactions 98<br />
3.8 Conclusions 100<br />
List of Abbreviations 101<br />
List of Symbols 101<br />
References 102<br />
Atomic Force Microscopy in Process Engineering 81<br />
© 2009, Elsevier Ltd
82 3. QUANTIFICATION OF PARTICLE–BUBBLE INTERACTIONs<br />
3.1 IntroDuCtIon<br />
The attachment of particles to bubbles in solution is of fundamental<br />
importance to several industrial processes most notably in froth flotation.<br />
Froth flotation is a significant industrial process, used primarily in the<br />
separation of mineral particles <strong>and</strong> also in the treatment of wastewater.<br />
The process of flotation involves the suspension of finely ground mineral<br />
particles in a chamber through which large volumes of air or another<br />
gas are bubbled. Hydrophobic particles will attach more readily to air<br />
bubbles passing through the medium <strong>and</strong> will thus rise to the top of<br />
the chamber where they form froth at the surface <strong>and</strong> become separated<br />
from other materials. Hydrophilic particles, which are less able to attach<br />
to the air bubbles, will eventually sink to the bottom of the chamber. In<br />
addition, a number of additives, including surfactants, termed collectors,<br />
may also be introduced to increase the hydrophobicity of the particles of<br />
interest <strong>and</strong> consequently increase the efficiency or specificity of the flotation<br />
process [1, 2].<br />
It follows that an underst<strong>and</strong>ing of the nature <strong>and</strong> strength of the interactions<br />
between colloidal particles suspended in solution <strong>and</strong> air bubbles<br />
is of fundamental importance to developing new ways of increasing flotation<br />
efficiency <strong>and</strong> modulating specificity. Atomic force microscopy<br />
is one technique which is able to quantitatively measure interactions<br />
between single particles <strong>and</strong> interfacial boundaries. Over the past decade<br />
the atomic force microscope (AFM) has been adapted for use in studying<br />
the forces involved in the attachment of single particles to bubbles in<br />
the laboratory. This allows the measurement of actual Derjaguin, L<strong>and</strong>au,<br />
Vervey <strong>and</strong> Overbeek (DLVO) forces <strong>and</strong> hydrodynamic, hydrophobic<br />
<strong>and</strong> adhesive contacts to be measured under different conditions. In<br />
addition, contact angles may be calculated from features of force versus<br />
distance curves. The purpose of this chapter is to illustrate how the AFM<br />
<strong>and</strong> particularly the colloid probe technique can be used to make measurements<br />
of single particle–bubble interactions <strong>and</strong> to summarise the<br />
current literature describing such experiments.<br />
3.2 PArtICLE–BuBBLE IntErACtIonS<br />
During collisions between mineral particles <strong>and</strong> air bubbles in flotation<br />
cells, attachment is determined by the thinning of a film of water in the<br />
intervening space. This consists of a layer of bulk water <strong>and</strong> a hydration<br />
layer which surround both the particle <strong>and</strong> the bubble (see Figure 3.1). On<br />
initial approach, the bulk fluid layer will become displaced [3]. However,<br />
the time required for this bulk layer to drain from the confined space
Bubble<br />
3.2 PARTICLE–BUBBLE INTERACTIONs 83<br />
Hydration<br />
layers<br />
between the particle <strong>and</strong> the bubble will add a hydrodynamic component<br />
to the attachment kinetics. In flotation, if the time required for this film to<br />
rupture (the induction time) is less than the actual contact time, then the<br />
particle will be unable to attach to the bubble [4]. Further approach will<br />
cause the hydration layers to become thinner <strong>and</strong> less stable. As this layer<br />
becomes destabilised, the particle <strong>and</strong> bubble will be allowed to attach<br />
directly, leading to the formation of a three phase contact (TPC) line. The<br />
more hydrophobic the particle surface, the thinner <strong>and</strong> less stable any<br />
hydration layer will be. Consequently, the more hydrophobic in character<br />
a particle is, the more readily it will attach to air bubbles in solution. This<br />
is the mechanism by which the flotation of fine particles is achieved [1, 2].<br />
Derjaguin <strong>and</strong> Dukhin [5] explain that the main thermodynamic parameter<br />
involved in the particle–bubble interaction is the disjoining pressure,<br />
which is defined by them as the derivative of the free energy with respect<br />
to the thickness h of this wetting layer per unit area. This disjoining pressure<br />
is formed by a combination of pressures:<br />
P � N � A � S<br />
( h) ( h) ( h) ( h)<br />
Particle<br />
Bulk fluid<br />
layer<br />
FIGurE 3.1 Simple illustration of a basic particle–bubble interaction. On approach<br />
there are three fluid layers present between the particle <strong>and</strong> the bubble. These are hydration<br />
layers on both the particle <strong>and</strong> the bubble (represented by the dashed line) <strong>and</strong> a layer<br />
of bulk fluid in the middle. The confinement <strong>and</strong> drainage of this bulk fluid layer leads to a<br />
hydrodynamic component to the interactions.<br />
(3.1)<br />
where P is the disjoining pressure, N the ionic double layer repulsion<br />
per unit area, A the contribution due to van der Waals interactions per<br />
unit area <strong>and</strong> S (h) the contribution of particle surface hydrophobicity to<br />
the disjoining pressure. S (h) will tend towards zero for very hydrophobic<br />
surfaces.<br />
As described by Sutherl<strong>and</strong> [6, 7], the probability of a particle being<br />
collected during flotation (P) is a composite of the probability of particle–<br />
bubble collision (P c), the probability of such a collision leading to adhesion
84 3. QUANTIFICATION OF PARTICLE–BUBBLE INTERACTIONs<br />
(P a) <strong>and</strong> the probability that a particle–bubble aggregate will become<br />
detached (P d):<br />
P � P P ( 1 � P )<br />
c a d<br />
(3.2)<br />
When conducting a bubble–particle interaction measurement using an<br />
AFM or similar piece of equipment, the probability of collision is determined<br />
by the user, <strong>and</strong> as such will not be considered further here. In<br />
terms of kinetics the attachment of a particle to the bubble depends upon<br />
the overcoming of an energy barrier by the kinetic energy imparted during<br />
the collision [6, 8]:<br />
⎛ E ⎞<br />
1<br />
Pa<br />
� exp ⎜<br />
⎜�<br />
⎝⎜<br />
E ⎠⎟<br />
k<br />
(3.3)<br />
where E 1 is the energy barrier for adhesion <strong>and</strong> E k the kinetic energy of the<br />
collision. The energy barrier is a result of the interaction of the repulsive<br />
<strong>and</strong> attractive forces acting between the particle <strong>and</strong> the bubble. The sum of<br />
forces between the particle <strong>and</strong> bubble can be summarised as follows:<br />
F � Fd � Fe � Fh<br />
(3.4)<br />
where F d <strong>and</strong> F e are the London dispersion van der Waals force <strong>and</strong> the<br />
electrical double layer force, respectively, representing traditional DLVO<br />
theory forces; <strong>and</strong> F h the hydrophobic attractive force between the particle<br />
<strong>and</strong> the bubble. When conducting AFM measurements between<br />
colloidal particles <strong>and</strong> a bubble, these forces may be manifested in the<br />
approach part of the force–distance curve. Models which describe longrange<br />
forces may be fitted to force–distance data to investigate the nature<br />
of the interaction under observation.<br />
Finally, the probability of detachment can be described by the following<br />
relationship [8]:<br />
P<br />
d<br />
Wa E<br />
� �<br />
E<br />
� ⎛ ⎞<br />
1<br />
exp ⎜ ’<br />
⎝⎜<br />
⎠⎟<br />
k<br />
(3.5)<br />
Here E k ’ is the kinetic energy of detachment <strong>and</strong> is not to be confused<br />
with E k the kinetic energy of collision; W a the work of adhesion. This latter<br />
quantity W a is of interest when carrying out AFM experiments as it<br />
can be directly related to adhesion measurements carried out. According<br />
to Johnson–Kendall–Roberts (JKR) theory, this is related to adhesion by<br />
the following association [9]:<br />
F<br />
ad<br />
RW<br />
� 3π<br />
2<br />
a<br />
(3.6)
3.2 PARTICLE–BUBBLE INTERACTIONs 85<br />
where F ad is the measured adhesion force <strong>and</strong> R the particle radius of<br />
curvature. This relationship assumes an idealised geometry of a sphere<br />
approaching a flat surface. If the air bubble is much greater in size than<br />
the particle, then this approximation is valid. However, it should also<br />
be noted that in practice the particles present during flotation are most<br />
likely to be irregular in shape, far removed from the idealised sphere.<br />
The first reported use of the AFM to measure interaction forces between<br />
colloidal particles <strong>and</strong> air bubbles in a fluid environment was by Butt<br />
[10], who measured the interaction between glass beads <strong>and</strong> an air bubble<br />
in water <strong>and</strong> between glass beads <strong>and</strong> water droplets in air. This was<br />
shortly followed by another study carried out by Ducker et al. [11]. In the<br />
work of Butt, air bubbles were immobilised onto the bottom of a polytetrafluoroethylene<br />
(PTFE) cuvette. It was concluded that hydrophilic particles<br />
approaching air bubbles experienced repulsive forces on approach.<br />
Conversely hydrophobic particles snapped in to the bubbles upon making<br />
close approach, becoming trapped within the air–water interface.<br />
In other studies reported in the literature, various immobilisation strategies<br />
have been used including highly ordered pyrolytic graphite (HOPG)<br />
surfaces [12]. When a hydrophobic substrate is used, such as HOPG, in<br />
an aqueous environment, the bubble will try to minimise its area of contact<br />
with the water by attaching itself to the substrate. As a result hydrophobic<br />
substrates are more suitable surfaces to attach the bubbles to than<br />
hydrophilic surfaces. Ducker et al. [11] used a slightly more sophisticated<br />
approach. Here a thin layer of mica with a small perforation was placed<br />
over a HOPG substrate. The bubble was then placed over the hole. This<br />
served to clamp the bubble in place, minimising any potential lateral<br />
movement of the bubble during measurement acquisition. Other immobilisation<br />
strategies involve attaching the bubble to a scratch made on the<br />
inside surface of a Petri dish [13] <strong>and</strong> attaching the bubbles to hydrophobic<br />
squares created by functionalising a surface with octanethiol [14].<br />
Producing air bubbles of an appropriate size using a small micro-syringe<br />
or micro-pipette <strong>and</strong> laboratory air is a relatively simple operation. To create<br />
a bubble of 250-�m diameter would require approximately 0.033 �l of air,<br />
assuming the bubble was hemispherical in shape. Creating bubbles of a size<br />
much smaller than this size becomes problematic. As the size of the bubble<br />
decreases, the Laplace pressure (the pressure difference between that inside<br />
the bubble <strong>and</strong> that outside the bubble) will increase. This makes small<br />
bubbles unstable <strong>and</strong> short lived as this high pressure will increase their<br />
tendency to dissolve in the surrounding fluid. However, there have recently<br />
been reports of very small nano-scale bubbles existing on hydrophobic<br />
interfaces under the right conditions [15–18]. It has been suggested that<br />
these nano-bubbles are responsible for the hydrophobic attraction observed<br />
between hydrophobic surfaces [15, 16, 19]. Jump-in between hydrophobic<br />
surfaces would thus be expected to occur when nano-bubbles attached to<br />
opposite hydrophobic surfaces come into contact.
86 3. QUANTIFICATION OF PARTICLE–BUBBLE INTERACTIONs<br />
3.3 DEtErmInAtIon oF PArtICLE–BuBBLE<br />
SEPArAtIon<br />
One of the main uncertainties involved when making single particle<br />
bubble measurements is the determination of the separation distance<br />
between the probe <strong>and</strong> the air–liquid interface. With a contact between<br />
hard surfaces, there is normally a clear transition as contact is made, but<br />
for deformable surfaces this is not necessarily obvious. First, deformation<br />
of the bubble by both long-range interaction forces prior to contact <strong>and</strong><br />
compression by the probe after contact needs to be taken into account<br />
(Figure 3.2). When a hydrophilic particle comes into contact with a bubble,<br />
a thin wetting film may separate the particle from the bubble. As the<br />
particle encounters this wetting film at small separations, the air–liquid<br />
interface will become deformed due to hydrodynamic factors, producing<br />
an apparent repulsion [20]. A hydrophobic particle may sit partially<br />
inside the bubble to a distance dependent upon their contact angle, or<br />
may be completely engulfed within the bubble. Together, these factors<br />
make an exact determination of the separation distance <strong>and</strong> attachment<br />
point more problematic than when considering interactions between<br />
hard solid surfaces.<br />
In the early paper by Ducker et al. [11], it was assumed that the bubble<br />
was deformed in a linear manner, resembling a Hookean spring. As such<br />
d 0 d n<br />
FIGurE 3.2 On approach of a particle into the vicinity of the bubble, long-range interactions<br />
will result in deformation of the bubble prior to contact. The nominal distance (d n) is the<br />
distance that would exist between the particle <strong>and</strong> the bubble if deformation did not occur.<br />
The actual distance (d 0) differs from d n by the size of the deformation. The size of the particle<br />
here is important as the magnitude of interaction forces will scale with its radius (R p), if the<br />
particle is a sphere. Note that in the illustration here the bubble is deforming as from a net<br />
repulsive interaction. With a net attraction deformation will be towards the particle.<br />
R p
3.3 DETERMINATION OF PARTICLE–BUBBLE sEPARATION 87<br />
the stiffness of the total system will behave as for any two linear springs<br />
in series:<br />
1 1 1<br />
� �<br />
ktot kc kbubble<br />
(3.7)<br />
where k b, k c <strong>and</strong> k tot are, respectively, the spring constants of the bubble,<br />
the cantilever <strong>and</strong> the bubble <strong>and</strong> lever combined. This means that the<br />
probe pressing against the bubble would lead to a gradient in the contact<br />
region based on both the deflection of the lever <strong>and</strong> the ‘stiffness’ of the<br />
bubble. The contact slope should thus be [21]:<br />
∆x<br />
k<br />
�<br />
∆z<br />
k � k<br />
bubble<br />
c tot<br />
(3.8)<br />
where �x <strong>and</strong> �z are the changes in cantilever deflection <strong>and</strong> piezotranslation<br />
distance, respectively. It is also assumed that the particle<br />
approaches in a direction perpendicular to the interface. If this is not the<br />
case, then the interaction becomes somewhat more complex due to the<br />
potential slippage of the particle along the interface.<br />
Attard <strong>and</strong> Miklavic [21, 22] in a thorough theoretical treatment of<br />
bubble deformation concluded that for small deformations relative to<br />
the bubble radius, air bubbles behave like linear Hookean springs, with<br />
the deformation given by equation (1.1) in Chapter 1. Unlike with solids,<br />
where the material properties determine stiffness, it is the interfacial<br />
tension <strong>and</strong> the pressure drop across the interface which determine the<br />
stiffness of the bubble. The same behaviour is also displayed by liquid<br />
droplets. Interestingly, when a micro-manipulation rig was used to apply<br />
large deformations (i.e. the deformation was �30% of the bubble diameter)<br />
to air bubbles in aqueous solutions, they were found to have a pseudoelastic<br />
behaviour [23]. However, deformation induced during AFM-based<br />
measurements is orders of magnitude smaller than this. Currently it is<br />
uncertain as to how large a deformation needs to occur to move from a<br />
linear, Hookean, deformation to a pseudo-elastic deformation.<br />
To calculate the separation distance, this linear deformation needs to<br />
be taken into account. From the slope <strong>and</strong> intercept of the contact part of<br />
the force curve, this can be estimated [12]:<br />
x<br />
d � �z � d<br />
� �<br />
c<br />
�<br />
c<br />
c<br />
(3.9)<br />
where � c is the slope of the contact region on the force curve (� c � �x/�z)<br />
<strong>and</strong> c the intercept; d c the particle–bubble distance when in ‘contact’,<br />
i.e. the thickness of any wetting film, etc. or the depth to which the interface
88 3. QUANTIFICATION OF PARTICLE–BUBBLE INTERACTIONs<br />
has been penetrated. This still leaves the need to either obtain the thickness<br />
of the wetting film from another source or assume that it is too thin to<br />
be significant, which may not be the case, particularly for very hydrophilic<br />
particles.<br />
Perhaps the simplest method for finding the zero contact point is to use<br />
the point at which the probe snaps-in during a jump to contact as suggested<br />
by Butt [9]. In cases where there are only repulsive forces evident prior to<br />
contact, there will be no snap-in <strong>and</strong> this method cannot be used. However,<br />
in cases where estimates of particle contact angle are to be made from force<br />
curves, this approach can be very useful. An interesting approach has been<br />
described by Gillies et al. [24]. At large separations, interaction forces are<br />
very weak, <strong>and</strong> thus the bubble or droplet will behave as though it is rigid<br />
at such separations. The force versus distance data can be shifted along<br />
the distance axis to coincide with the linear Poisson–Boltzmann theory for<br />
rigid bodies. However, this technique requires determination of the surface<br />
potentials from some other source, such as by electrophoresis, making its<br />
implementation problematic in laboratories where the appropriate equipment<br />
is not available.<br />
3.4 DEtErmInAtIon oF ContACt AnGLE From<br />
ForCE–DIStAnCE CurvES<br />
As mentioned previously, the attachment of particles to air bubbles in<br />
an aqueous environment is largely mediated by the degree of hydrophobicity<br />
of the particle. A hydrophobic particle will prefer to be in contact<br />
with the bubble, minimising its contact with surrounding water, whereas<br />
a hydrophilic particle will retain a thin wetting film. One conventional<br />
method of measuring the wettability, <strong>and</strong> hence the degree of hydrophobicity<br />
of a surface is to measure the contact angle of a drop of water on<br />
that surface, which is the angle formed by the TPC line [25–27]. When<br />
the particle comes into contact with an air bubble, the TPC formed will<br />
likewise also contain a contact angle.<br />
Figure 3.3 shows a basic schematic representation for the particle–<br />
bubble interaction. The particle has penetrated the bubble to a distance D.<br />
The angle � represents the immersion angle of the particle, which gives<br />
an indication of the position of the TPC line with regard to the particle.<br />
The contact angle is indicated by � <strong>and</strong> corresponds to the angle of contact<br />
internal to the droplet during conventional contact angle measurements.<br />
The contact angle is related to the interfacial tensions of the three<br />
interfaces present around the TPC by the Young equation:<br />
� � �<br />
cos� �<br />
�<br />
SV SL<br />
LV<br />
(3.10)
3.4 DETERMINATION OF CONTACT ANgLE FROM FORCE–DIsTANCE CURvEs 89<br />
θ<br />
FIGurE 3.3 A particle in direct contact with a bubble with a net force of zero, at which<br />
point it is immersed to a distance D. When no net force is present then the immersion angle<br />
� is equal to the contact angle �. The contact angle may be receding or advancing depending<br />
upon the direction of travel of the particle.<br />
where � indicates interfacial tension, with the subscripts SV, SL <strong>and</strong> LV<br />
denoting the solid–vapour, solid–liquid <strong>and</strong> liquid–vapour interactions,<br />
respectively.<br />
In addition, the tension around the TPC line, called the line tension,<br />
may need to be taken into account. For macro-scale experiments, the line<br />
tension is negligible <strong>and</strong> is usually ignored, with the Young equation<br />
being a good approximation. However, at small length scales, the line<br />
tension becomes more significant <strong>and</strong> further terms need to be added.<br />
For a spherical particle in an interface, the contact angle will be [28, 29]:<br />
� � �<br />
cos� �<br />
�� � �<br />
SL SV<br />
LV<br />
(3.11)<br />
where � is the line tension <strong>and</strong> � the geodesic curvature of the TPC line.<br />
As the TPC line is circular, then:<br />
�<br />
�<br />
�<br />
1 r sin<br />
(3.12)<br />
where r is the radius of the circle drawn out by the TPC line (equal to the<br />
radius of the particle when immersed halfway).<br />
In many systems, there exists a contact angle hysteresis between an<br />
advancing (� a) <strong>and</strong> a receding contact angle (� r), i.e. a different contact angle<br />
may be measured depending upon whether the TPC line is advancing or<br />
receding over a particle surface. The reason for this hysteresis is generally<br />
ascribed to roughness <strong>and</strong> heterogeneity of the surface chemistry, although<br />
other factors may be of importance, such as the existence of a contaminant<br />
layer or polymer coatings, etc. [30–35].<br />
As the particle enters into the bubble, the TPC line is receding across the<br />
particle surface. As a result it is the receding contact angle which may be calculated<br />
from the approach part of the force curve, <strong>and</strong> the advancing contact<br />
α<br />
D
90 3. QUANTIFICATION OF PARTICLE–BUBBLE INTERACTIONs<br />
angle when the particle is being extricated from the bubble. The jump-<br />
in force is dominated by capillary forces, at least in the case of a particle with<br />
a hydrophobic enough surface to be able to form a TPC line. As such, the<br />
capillary force may be related directly to � r, as illustrated by equation (3.15)<br />
[36–38]:<br />
F � 2πR� sin� sin( � � �)<br />
CAP r<br />
(3.13)<br />
where � represents the interfacial tension. When the net forces are zero<br />
(i.e. F CAP � 0), � r will be equal to �; when this occurs, equation (1.14) may<br />
be used to calculate � r [36, 39]:<br />
cos<br />
� r<br />
( R Dr<br />
)<br />
�<br />
R<br />
�<br />
(3.14)<br />
Here the penetration depth D r can be extracted from force curves by<br />
taking the distance from the initial jump-to-contact to the point at which<br />
the force is zero, i.e. there is no net force acting on the cantilever. This is<br />
illustrated in Figure 3.4. When the particle is withdrawn from the bubble,<br />
the TPC line is advancing across the particle, hence contact angles determined<br />
are called the advancing contact angle (� a). If adhesion occurs,<br />
D a<br />
Dr<br />
FIGurE 3.4 A force curve, showing the relevant measurements to obtain D r <strong>and</strong> D a. D r<br />
is found by measuring the distance from jump-in to the point at which a net force of zero,<br />
equivalent in size to the free level force, is reached. From this distance on the approach<br />
curve, the receding contact angle � r is obtained. From a similar measurement on the retract<br />
curve (D a) the advancing contact angle � a may be obtained.
3.5 EFFECT OF sURFACE PREPARATION ON PARTICLE–BUBBLE INTERACTIONs 91<br />
then the advancing contact angle � a may be obtained from the retract<br />
part of the force trace. Here the adhesion force F AD can be related to the<br />
advancing contact angle:<br />
2<br />
a<br />
2 �<br />
FAD � 2πR�<br />
sin<br />
2<br />
(3.15)<br />
Other methods have been developed to measure contact angles using<br />
the AFM in tapping mode. Pompe et al. [40] scanned liquid drops on surfaces.<br />
By then using a cross section of the topography, they measured the<br />
contact angle <strong>and</strong> three-phase line tension parameters for the liquid drops.<br />
3.5 EFFECt oF SurFACE PrEPArAtIon on<br />
PArtICLE–BuBBLE IntErACtIonS<br />
3.5.1 Effect of Particle Surface Chemistry on Particle–Bubble<br />
Interactions<br />
As would be expected, the surface chemistry of the colloid probes has<br />
an important effect on their attachment to bubbles in aqueous solutions.<br />
In the literature there have been a number of studies in which the effect<br />
of the surface chemistry of various probes has been investigated for their<br />
importance in particle attachment to bubbles.<br />
In the first set of AFM-based experiments to measure particle–bubble<br />
interactions in the literature, as described by Butt [10], untreated hydrophilic<br />
glass particles were allowed to interact with air bubbles in aqueous<br />
media. When the glass approached the surface, a linear repulsive force<br />
was measured on contact, with no jump-in prior to contact. This was<br />
echoed in the work by Fielden et al. using hydrophilic silica particles [37],<br />
who also noted a linear repulsion with no jump-in. This is hardly surprising<br />
as both silica <strong>and</strong> glass have a tendency to carry a negative charge in<br />
water due to Si–OH groups on the surface [41], <strong>and</strong> the air–water interface<br />
of air bubbles also tends to be negatively charged for the majority of<br />
pH values, as evidenced by �-potential measurements [42, 43], leading to<br />
repulsive electrical double layer forces. In addition it would be expected<br />
that van der Waals forces between silica <strong>and</strong> air in water would also be<br />
repulsive, due to a negative Hamaker constant for the interaction of silica<br />
<strong>and</strong> air across water [26, 44]. This means that the DLVO forces in general<br />
will all tend towards repulsion in this case.<br />
Interestingly, a different result was reported by Ducker et al. [11] when<br />
interacting a silica sphere with an air bubble in water. Jump-in events<br />
were observed prior to contact, at a distance of 50 nm, before being followed<br />
by a linear repulsion after contact was made. It was speculated by<br />
the authors that at small separations, the air–water interface may change
92 3. QUANTIFICATION OF PARTICLE–BUBBLE INTERACTIONs<br />
sign under the influence of the approaching silica particle. A similar argument<br />
was used to explain the adhesion forces observed (�4 mN m �1 ),<br />
measured during retraction of the hydrophilic silica particles by Fielden<br />
et al. It has been noted elsewhere that if there is a sufficient difference<br />
in the magnitude of the surface potentials of two approaching bodies,<br />
then even if the bodies have potentials of the same sign at sufficiently<br />
small separations, an attractive force may occur [45]. It is worth bearing<br />
in mind that due to the very small surface areas of the interactions, these<br />
types of experiments are very sensitive to any contamination, even if great<br />
care is taken to prevent this. Deviations from the expected behaviour<br />
could easily be caused as a result of such contamination. In the literature,<br />
papers describing colloid probe-based particle–bubble interactions<br />
describe stringent techniques to minimise the risk of contamination. The<br />
potential problems due to probe contamination are a serious concern for<br />
all SPM techniques. See Chapter 1 for more information on this subject.<br />
Glass particles used by Butt et al. [10] were made hydrophobic by<br />
silanizing them in an atmosphere of dichloromethane. When these particles<br />
were allowed to approach an air bubble, their behaviour was found to<br />
have altered from that of hydrophilic particles. This time jump-in events<br />
occurred on approach to the bubble surface with no repulsive interaction<br />
prior to contact. When the particle was retracted large adhesion occurred,<br />
with the cantilever deflecting by a greater amount than could be detected<br />
by the instrument being used [10]. In the study by Fielden et al. [37] the<br />
silica particles were hydrophobized by either reacting with octadecyltrichlorosilane,<br />
or by dehydroxylation to create a surface with a more moderate<br />
degree of hydrophobicity. For the two types of hydrophobic surface,<br />
interactions between the particles <strong>and</strong> the bubbles were attractive at small<br />
separation distances. Furthermore, the attraction was dependent upon<br />
the degree of hydrophobicity of the surface. Frequently the particle was<br />
engulfed by the air bubble <strong>and</strong> much larger adhesion forces were measured<br />
than with the hydrophilic particle. It was concluded that the jumpin<br />
events observed were a result of hydrophobic attraction. These two<br />
studies illustrate the importance of hydrophobicity in particle–bubble<br />
attachment <strong>and</strong> for the degree of stability of particle–bubble aggregates.<br />
Preuss <strong>and</strong> Butt [39] measured the interactions of bubbles with silica<br />
particles coated with different surface concentrations of alkyl-thiol<br />
molecules <strong>and</strong> the resultant receding contact angles both with the colloid<br />
probe technique <strong>and</strong> by conventional means. They noticed that as<br />
well as more hydrophilic particles having shorter jump-in distances<br />
<strong>and</strong> less adhesion than hydrophobic particles, both the � r <strong>and</strong> adhesion<br />
forces increased as the mole fraction of alkyl-thiols was increased.<br />
As an increase in contact angle is conventionally related to an increase<br />
in hydrophobicity of a surface [27], the relationship between hydrophobicity<br />
<strong>and</strong> the stability of particle bubble attachment is again reiterated.
3.5 EFFECT OF sURFACE PREPARATION ON PARTICLE–BUBBLE INTERACTIONs 93<br />
In addition the authors compared the contact angles obtained from the<br />
AFM measurements with those obtained conventionally on flat surfaces.<br />
It was found that the � r values obtained by the two methods differed,<br />
depending upon the contact angles measured. At low contact angles (i.e.<br />
more hydrophilic surfaces), contact angles measured with the colloid<br />
probe were higher than those measured on planar surfaces. At contact<br />
angles �60° the values obtained with the colloid probes were lower than<br />
those measured on flat surfaces, with measurements at intermediate contact<br />
angles in close agreement. These differences were explained by the<br />
authors as being possibly due to the differences in line tension between<br />
the different experimental set-ups [39].<br />
The interactions of ZnS spheres with air bubbles at a range of solution<br />
pH values was examined by Gillies et al. [46, 47]. On approach, repulsion<br />
between the spheres <strong>and</strong> the bubbles was observed due to DLVO forces (the<br />
micro-spheres have a negatively charged surface in solution under the conditions<br />
of the experiments) <strong>and</strong> confinement of hydration layers, prior to a<br />
jump-in event. The magnitude of the repulsive force prior to jump-in was<br />
found not to vary upon alteration of solution pH to a significant effect at<br />
pH values of 8.5 or less. On increasing the pH above this level, the repulsion<br />
was increased by an order of magnitude, due to a much greater concentration<br />
of negative charges on the micro-sphere surface. Contact angles measured<br />
at the same time appeared to alter, based upon the number of times<br />
the micro-sphere had previously interacted with the bubble surface, leading<br />
to speculation that the ZnS surface was changing, either by being cleaned<br />
or coated by the interaction. In addition a slight increase in the receding<br />
contact angle was observed concomitant with a rise in pH. Leaving the<br />
micro-spheres in zinc solutions for extended periods of time caused the<br />
particle surface to change from a predominantly zinc hydroxide to a predominantly<br />
zinc oxide surface, resulting in a greater degree of hydrophobicity.<br />
This resulted in contact angles measured in these aged micro-spheres<br />
which were approximately 10° greater than those of the none-aged spheres.<br />
Wangsa-Wirawan et al. [13] measured the adhesion forces due to the<br />
interactions between protein inclusion bodies <strong>and</strong> air bubbles. Both pH<br />
<strong>and</strong> ionic strength of the surrounding solution was altered <strong>and</strong> the effects<br />
measured. A maxima in the adhesion forces was reached at pH 5, with a<br />
decrease at higher pH values, although adhesion was still observed. The<br />
inclusion bodies were expected to carry a net overall negative charge at<br />
pH values �5, leading to electrostatic repulsion between the bodies <strong>and</strong><br />
the negatively charged air–water interface. It was concluded that whilst<br />
the electrostatic forces modulated the adhesion, hydrophobic interactions<br />
between the inclusion bodies <strong>and</strong> the air–bubble played a more dominant<br />
role in the adhesion. Additionally it was reported that the ionic strength<br />
had an effect on the measured adhesion values in a way that could not be<br />
explained purely by DLVO interactions.
94 3. QUANTIFICATION OF PARTICLE–BUBBLE INTERACTIONs<br />
3.5.2 Effect of Surfactant on Particle–Bubble Interactions<br />
During the froth flotation process, surfactant compounds are commonly<br />
added to the flotation chamber to modify bubble–particle<br />
attachment <strong>and</strong> increase mineral recovery. A number of AFM-based<br />
measurements have been carried out to assess the effect of surfactants in<br />
solution on particle–bubble interactions <strong>and</strong> hence, presumably, on flotation<br />
efficiency.<br />
Preuss <strong>and</strong> Butt [36, 48] studied the effects of adding two surfactants,<br />
sodium dodecyl sulphate (SDS) <strong>and</strong> dodecyltrimethylammonium bromide<br />
(DTAB), to solution when carrying out measurements between<br />
both hydrophobic <strong>and</strong> hydrophilic particles with air bubbles. When<br />
hydrophilic silica particles were allowed to approach the bubble in aqueous<br />
solution without surfactant present, repulsive forces were measured<br />
prior to contact, as reported previously for hydrophilic silica particles [37,<br />
39]. As the concentration of SDS in solution was increased, this repulsive<br />
force also increased <strong>and</strong> the decay length became decreased. As SDS has<br />
a negative charge in solution, as do the silica–water <strong>and</strong> air–water interfaces,<br />
the most likely explanation was that the repulsion was electrostatic<br />
in origin <strong>and</strong> increased with increasing quantities of SDS at the interfaces.<br />
When the effects of DTAB were investigated long-range forces between<br />
the particles <strong>and</strong> the bubbles occurred, <strong>and</strong> adhesion was observed when<br />
retracting the particle, most probably due to DTAB coating the silica surface<br />
<strong>and</strong> increasing its hydrophobicity.<br />
When silanized hydrophobic silica particles were used, capillary forces<br />
were dominant, with large adhesions on the order of 70 mN m �1 detected<br />
upon particle retraction [36, 48]. In the absence of SDS, small repulsive<br />
forces were measured on approach, probably due to electrostatic repulsion.<br />
When SDS was introduced, adhesion appeared to be dependent<br />
upon the maximum force used to press the particle into the bubble, reaching<br />
a threshold value of approximately 6 mN m �1 for an SDS concentration<br />
of 5.6 mM. Below this threshold value no adhesion was seen, with no<br />
hysteresis between the approach <strong>and</strong> the retract parts of the force curves,<br />
suggesting that the SDS was reducing the interaction between the particle<br />
<strong>and</strong> the bubble. As the loading force increased above the threshold, a<br />
jump-in event occurred <strong>and</strong> high adhesion was observed. This threshold<br />
force was increased, as the SDS concentration was increased as illustrated<br />
in Figure 3.5. When the effects of DTAB were investigated, they were<br />
found to follow a similar trend to that seen with SDS. At concentrations<br />
�6 or 12 mM, for SDS or DTAB, respectively, formation of the TPC was no<br />
longer observed for the load forces that were reached [36, 48]. The most<br />
likely explanation for this behaviour is that a surfactant film was built up<br />
between particle <strong>and</strong> bubble, preventing TPC formation. As the surfactant<br />
concentration was increased, this film would most likely be thicker <strong>and</strong>
3.5 EFFECT OF sURFACE PREPARATION ON PARTICLE–BUBBLE INTERACTIONs 95<br />
Fthr/R/mN/m<br />
8<br />
6<br />
4<br />
2<br />
0<br />
SDS<br />
DTAB<br />
0 2 4 6 8 10 12<br />
Surf.conc./mM<br />
FIGurE 3.5 Effect of SDS <strong>and</strong> DTAB concentration on the (normalised) threshold force<br />
(F/R) needed to achieve a TPC with an air bubble in aqueous solution. Reprinted with permission<br />
from [36]. Copyright 1998 American Chemical Society.<br />
more stable. To make a TPC contact, the loading force would have to be<br />
sufficient to cause penetration of the surfactant film, characterised by the<br />
threshold force, which would increase with thicker surfactant films.<br />
Recently, force interaction measurements have been carried out<br />
between silica glass spheres <strong>and</strong> colloidal gas aphrons (CGAs) [49].<br />
CGAs are a form of very small (�100 �m) surfactant-stabilized bubbles<br />
created by high shear rate flows, first described by Sebba in 1971 as<br />
micro-foams [50]. The use of aphrons has been proposed for a number<br />
of potentially useful applications. These include the flotation of fine<br />
mineral particles [51, 52], the treatment of particulates <strong>and</strong> contaminant<br />
chemicals including solvents <strong>and</strong> heavy metals from wastewater [53–56],<br />
treatment of soil contamination [53, 57, 58], the harvesting of microbial<br />
cells [59] <strong>and</strong> extraction of protein products [58, 60, 61], <strong>and</strong> the production<br />
of tissue engineering scaffolds [62]. Sebba proposed a multi-walled<br />
structure for CGAs consisting of an outer surfactant bilayer, followed by<br />
an internal layer of surfactant solution, in turn separated by a surfactant<br />
monolayer from a central gas core. Although this putative structure has<br />
not been confirmed conclusively, there are a number of indirect experimental<br />
observations which suggest that the internal structure of CGAs<br />
are different from those observed with conventional foams, including xray<br />
diffraction, <strong>and</strong> transmission electron microscopy (TEM) [58, 63].
96 3. QUANTIFICATION OF PARTICLE–BUBBLE INTERACTIONs<br />
Force, F/R (mN/m)<br />
A<br />
Force, F/R (mN/m)<br />
B<br />
1<br />
0.8<br />
0.6<br />
0.4<br />
0.2<br />
–150 –100 –50<br />
0<br />
0<br />
–0.2<br />
50 100 150 200 250 300<br />
–0.4<br />
–0.6<br />
–0.8<br />
–1<br />
1<br />
0.8<br />
0.6<br />
0.4<br />
0.2<br />
–0.4<br />
–0.6<br />
–0.8<br />
–1<br />
z-Displacement (nm)<br />
–150 –100 –50<br />
0<br />
0<br />
–0.2<br />
50 100 150 200 250 300<br />
z-Displacement (nm)<br />
FIGurE 3.6 Normalised force versus z-displacement curves obtained between aphrons<br />
created using SDS (A) <strong>and</strong> DTAB (B) in water. Each force curve is an average of several<br />
force–distance cycles (SDS � 10; DTAB � 9). In both cases a jump-in event is observed.<br />
However, when the glass sphere approaches the anionic SDS-stabilised aphron, a repulsive<br />
force is observed prior to contact, most likely due to repulsive (negative) charge interactions.<br />
When an identical glass probe approaches DTAB-stabilised aphrons, only attractive<br />
forces are seen prior to jump-in.<br />
Aphrons created in surfactant solutions which were either anionic<br />
sodium dodecyl sulfate (SDS) or cationic dodecyl trimethylammonium<br />
bromide (DTAB) by mixing at high sheer rates (with rotor speeds<br />
� 5000 rpm) were immobilised on hydrophobic HOPG <strong>and</strong> rinsed with<br />
an excess of high purity de-ionised water [49] to remove excess surfactant<br />
from solution. Aphrons were made in solutions with surfactant<br />
concentrations of 1 <strong>and</strong> 2 g l �1 for SDS <strong>and</strong> DTAB, respectively. Colloid<br />
probes consisting of glass spheres were then allowed to approach the<br />
surface of the aphrons (estimated bubble diameters were � 50 �m) <strong>and</strong><br />
the resultant interaction forces were measured. Figure 3.6 shows example<br />
force measurements obtained when approaching aphrons produced
3.6 EFFECT OF LOADINg FORCE ON PARTICLE–BUBBLE INTERACTIONs 97<br />
using SDS <strong>and</strong> DTAB. In both cases probes became attached to the CGA<br />
surfaces. However, when approaching SDS CGAs, a repulsive force was<br />
observed prior to contact. This force was not observed with the DTAB<br />
aphrons. The difference is most likely due to charge repulsion effects,<br />
which is due to the interaction between the negatively charged glass surface<br />
<strong>and</strong> the negatively charged SDS. In the case of DTAB, which contains<br />
a positive charge in solution, only attractive forces are observed prior to<br />
contact. When silica was allowed to interact with CGAs in bulk flotation<br />
experiments, the silica was found to separate with the CGA fraction<br />
when using DTAB to create the CGAs, but a much smaller fraction was<br />
recovered when using SDS. The attachment of silica to CGAs in solution<br />
was modulated by the charge interactions between silica <strong>and</strong> the CGAs<br />
in solution, with repulsive charges preventing the attachment of the particles<br />
to SDS CGAs in solution.<br />
3.6 EFFECt oF LoADInG ForCE on<br />
PArtICLE–BuBBLE IntErACtIonS<br />
As has been seen above in the work by Preuss <strong>and</strong> Butt, in the presence<br />
of surfactants [36, 39, 48], sometimes a threshold force is required to form a<br />
TPC contact, but there are likely to be other effects of the loading force on<br />
particle–bubble interactions. As the loading force is increased, deformation<br />
of the bubble would be expected to increase, particularly if the TPC line<br />
does not move over the particle surface. For other deformable surfaces,<br />
such as soft polymer interfaces, the maximum loading force reached has<br />
been observed to have a positive effect on the adhesion force [64]. This is<br />
not surprising, as the area of contact between the probe <strong>and</strong> the surface<br />
will increase as the surface becomes increasingly deformed. At small loading<br />
forces, where the force F
98 3. QUANTIFICATION OF PARTICLE–BUBBLE INTERACTIONs<br />
3.7 EFFECt oF HyDroDynAmICS on<br />
PArtICLE–BuBBLE IntErACtIonS<br />
As all of the particle–bubble interaction measurements carried out<br />
using an AFM-based system are undertaken in a liquid environment, it<br />
is necessary to consider the implications of hydrodynamic forces on the<br />
interactions being measured. During the process of froth flotation, the<br />
particles <strong>and</strong> bubbles are circulated around the flotation chamber at some<br />
speed, <strong>and</strong> to relate what happens in this industrial process to what happens<br />
in AFM experiments, some measure of the effects of the speed of<br />
interaction between particles <strong>and</strong> bubbles necessarily has to be made.<br />
Nguyen <strong>and</strong> Evans [66] considered the hydrodynamic force acting on a<br />
sphere approaching a bubble in fluid. The hydrodynamic drag force F h<br />
on the particle follows the following relationship, dependent upon the<br />
separation distance d from the bubble surface:<br />
Fh RVf<br />
� �6π� 1<br />
(3.16)<br />
where � is the dynamic viscosity of the surrounding fluid, V the velocity<br />
with which the particle approaches the bubble surface (not to be confused<br />
with the AFM piezo-drive speed), <strong>and</strong> f 1 a correction factor which accounts<br />
for the deviation of the drag force from Stokes law. It is within this correction<br />
factor that the term for the separation distance appears <strong>and</strong> was derived by<br />
Nguyen <strong>and</strong> Evans to be approximated by:<br />
f<br />
1<br />
R<br />
� 1 �<br />
4d<br />
⎛ ⎡<br />
⎢ ⎞<br />
⎜<br />
⎢ ⎝⎜<br />
⎠⎟<br />
⎣⎢<br />
1<br />
0. 719⎤<br />
0. 719<br />
⎥<br />
⎦⎥<br />
(3.17)<br />
Deformation of the bubble as the approaching particle encounters the<br />
thin wetting film around the bubble will lead to a reduction of V at small<br />
separation distances [20] relative to the driving speed.<br />
Nguyen et al. [12] studied the effect of the approach speed upon the<br />
forces obtained when a spherical glass particle approached an air bubble<br />
surface. In Figure 3.7 the effect on the measured force of different piezotranslation<br />
speeds is demonstrated. As the approach speed was increased,<br />
the repulsive particle–bubble forces at short separation distances were<br />
also increased. This was possibly due to limitations on the ability of<br />
the thin water film between particle <strong>and</strong> bubble to drain in the shorter<br />
time frames imposed by the higher approach velocities. This demonstrates<br />
that hydrodynamic forces were acting as an additional repulsive<br />
force. At low piezo-approach speeds of 0.6 �m s �1 the hydrodynamic forces<br />
were negligible <strong>and</strong> instead surface forces were dominant. This effect has
3.7 EFFECT OF HyDRODyNAMICs ON PARTICLE–BUBBLE INTERACTIONs 99<br />
AFM Measured interaction force, F (nN)<br />
6<br />
4<br />
2<br />
0<br />
0<br />
Approach speeds (µm/s)<br />
0.6<br />
3.0<br />
5.9<br />
12.2<br />
17.8<br />
27.9<br />
39.1<br />
65.4<br />
97.8<br />
100 200 300 400 500 600<br />
Separation distance, D (nm)<br />
FIGurE 3.7 Effect of the speed of approach between a particle <strong>and</strong> bubble on the measured<br />
forces. Reprinted from [12]. Copyright 2004, with permission from Elsevier.<br />
also been characterised using the colloid probe technique for a sphere<br />
approaching a solid confining wall in a fluid environment [67–71]. Due to<br />
the change in deflection of the cantilever as a result of the action of forces<br />
upon it, the actual velocity experienced by the particle differs from that<br />
applied by the piezo. A plot of the actual velocity against measured force<br />
<strong>and</strong> calculated hydrodynamic force as determined by Nguyen et al. [12] is<br />
presented in Figure 3.8. As can be seen the velocity actually decreases as<br />
the particle–bubble separation distance is decreased due to the increase in<br />
repulsive force, causing the cantilever to become deflected upwards, away<br />
from the bubble surface. A similar effect was observed by Aston <strong>and</strong> Berg<br />
[72] when approaching a droplet of n-hexadecane with a glass sphere in an<br />
SDS solution. It was found that the oil–water interface became dimpled on<br />
approach, with the distance at which this occurred increasing with increasing<br />
approach speed.<br />
Another study by Nguyen et al. [38] examined the effect of the speed<br />
of approach of spherical polyethylene particles on the receding contact<br />
angles measured, obtained from force curves as described earlier. It was<br />
found that at relatively high approach speeds of �10 �m s �1 , there was<br />
a strong influence of the approach speed on measured contact angle.<br />
At speeds below this there was little influence on contact angle. It was<br />
surmised by the authors that at the lower speeds, the measurements<br />
represented a ‘static’ contact angle.
100 3. QUANTIFICATION OF PARTICLE–BUBBLE INTERACTIONs<br />
Normalised force, F/R (mN/m)<br />
1<br />
0.8<br />
0.6<br />
0.4<br />
0.2<br />
Velocity<br />
Force<br />
0<br />
0<br />
0 200 400 600 800 1000<br />
Separation distance, D (nm)<br />
FIGurE 3.8 The relative velocity V <strong>and</strong> the measured (points) <strong>and</strong> hydrodynamic<br />
forces (line), for a piezo-movement speed of 48.8 �m s �1 . Reprinted from [12], Copyright<br />
2004, with permission from Elsevier.<br />
3.8 ConCLuSIonS<br />
The AFM-based colloid probe technique has been used to investigate the<br />
interactions between single particles <strong>and</strong> air bubbles in aqueous solution<br />
with the aim of increasing the basic underst<strong>and</strong>ing of processes dependent<br />
upon this type of interaction. The effect of such particle properties as the<br />
degree of hydrophobicity has been examined in relation to properties such<br />
as particle–bubble adhesion, the favourability of long-range interaction<br />
forces to particle–bubble attachment, as well as measured contact angles<br />
in a number of studies. Measurements have also begun to characterise<br />
the effect of force <strong>and</strong> approach speed on these interactions. As froth flotation<br />
is a dynamic process, an underst<strong>and</strong>ing of the forces <strong>and</strong> kinetics<br />
involved in particle–bubble attachment may be necessary to marry AFMbased<br />
measurements to what is happening in the flotation chamber or<br />
other related industrial processes. The body of literature which now exists<br />
is not yet complete enough to give a full insight into how particle–bubble<br />
interactions on the single interaction level mediate the bulk process of<br />
froth flotation. In addition, measurements so far utilise micro-spheres with<br />
an idealised geometry. Particles used in mineral separation are more likely<br />
to be rough <strong>and</strong> irregular <strong>and</strong> quite likely more heterogeneous, leading to<br />
more complexity in their attachment to bubbles. In the industrial processes,<br />
there are a large number of different surfactants <strong>and</strong> other chemicals used<br />
to modulate mineral separation, depending upon the properties of the<br />
minerals of interest. In the AFM-based literature, only a small number of<br />
commonly available surfactants have been used. On the whole, the use of<br />
50<br />
40<br />
30<br />
20<br />
10<br />
Relative velocity, U (µm/s)
LIsT OF syMBOLs 101<br />
AFM-based techniques to study particle–bubble interactions shows a great<br />
amount of promise to elucidate the properties on the level of single interactions<br />
which determine effects observed at the macroscopic level; however,<br />
there is room for a great number nevertheless of further measurements <strong>and</strong><br />
experiments before a thorough underst<strong>and</strong>ing can be realised. However,<br />
the ability of the AFM <strong>and</strong> in particular the colloidal probe technique has<br />
a great deal of potential, due to its unique ability to measure interaction<br />
forces on such a small scale, to contribute greatly to the underst<strong>and</strong>ing of<br />
the underlying mechanisms in processes dependent upon the interactions<br />
between suspended colloids <strong>and</strong> deformable interfaces.<br />
LISt oF ABBrEvIAtIonS<br />
AFM Atomic force microscopy<br />
DLVO Derjaguin, L<strong>and</strong>au, Vervey <strong>and</strong> Overbeek<br />
DTAB Dodecyltrimethylammonium bromide<br />
HOPG Highly ordered pyrolytic graphite<br />
JKR Johnson, Kendall <strong>and</strong> Roberts theory<br />
PTFE Polytetrafluoroethylene<br />
SDS Sodium dodecyl sulphate<br />
TPC Three phase contact<br />
LISt oF SymBoLS<br />
c Intercept of force curve<br />
x Deflection m<br />
d Separation distance m<br />
D Penetration distance of particle into bubble m<br />
dc Particle bubble separation in contact m<br />
E1 Energy barrier for adhesion J<br />
Ek Kinetic energy of collision J<br />
’ Ek Kinetic energy of detachment J<br />
F Force N<br />
Fad Force of adhesion N<br />
FCAP Capillary force N<br />
Fd London dispersive van der Waals forces N
102 3. QUANTIFICATION OF PARTICLE–BUBBLE INTERACTIONs<br />
F e Electrical double layer forces N<br />
F h Hydrophobic force N<br />
k Spring constant N m �1<br />
k b Boltzmann’s constant (1.38 � 10 �23 ) J K �1<br />
k bubble Bubble spring constant N m �1<br />
k c Cantilever spring constant N m �1<br />
k m Uncorrected measured spring constant N m �1<br />
k ref Reference cantilever spring constant N m �1<br />
k tot Total spring constant of system N m �1<br />
P a<br />
P c<br />
P d<br />
Probability of adhesion<br />
Probability of collision<br />
Probability of detachment<br />
r Radius of TPC circle m<br />
R Radius of sphere m<br />
V Velocity of particle m s �1<br />
W Interaction energy J m �2<br />
W a Work of adhesion J<br />
z Distance travelled by z-piezo m<br />
� Immersion angle °<br />
� Interfacial tension N m �1<br />
� c<br />
Slope of contact region of force curve<br />
� hard Slope of contact region against hard surface nA m �1<br />
�l Distance of probe from end of cantilever m<br />
� Contact angle °<br />
� a Advancing contact angle °<br />
� r Receding contact angle °<br />
� Geodesic curvature of TPC<br />
� f Density of fluid Pa s<br />
� Line tension N m �1<br />
�i Imaginary component of hydrodynamic function<br />
references<br />
[1] J. Leja, Surface Chemistry of Froth Flotation, Plenum Press, New York, 1982.<br />
[2] B.A. Wills, Wills’ Mineral Processing Technology, seventh ed., Elsevier, Oxford, 2006.<br />
[3] V.I. Klassen, V.A. Mokrousov, An Introduction to the Theory of Flotation, second ed.,<br />
Butterworth & Co, London, 1963.
REFERENCEs 103<br />
[4] A.V. Nguyen, H. Stechemesser, Dynamics of the impact interaction between a fine<br />
solid sphere <strong>and</strong> a plane gas–liquid interface, in: D. Mobius, R. Miller (Eds.), Drops<br />
<strong>and</strong> Bubbles in Interfacial Research, Elsevier, Amsterdam, 1998, pp. 525–562.<br />
[5] B.V. Derjaguin, S.S. Dukhin, Theory of flotation of small <strong>and</strong> medium sized particles,<br />
Prog. Surf. Sci. 43 (1-4) (1993) 241–266.<br />
[6] R.-H. Yoon, Bubble–particle interactions in flotation, in: B.K. Parekh, J.D. Miller (Eds.),<br />
Advances in Flotation Technology, Society for Mining, Metallurgy, <strong>and</strong> Exploration,<br />
Inc, 1999, pp. 95–113.<br />
[7] K.L. Sutherl<strong>and</strong>, Physical chemistry of flotation. 11. Kinetics of the flotation process,<br />
J. Phys. Chem. 52 (1948) 394.<br />
[8] R.-H. Yoon, L. Mao, Application of extended DLVO theory, IV, J. Colloid Interface Sci.<br />
181 (1996) 613–626.<br />
[9] H.-J. Butt, B. Capella, M. Kapple, Force measurements with the atomic force microscope:<br />
technique, interpretation <strong>and</strong> applications, Surf. Sci. Rep. 59 (2005) 1–152.<br />
[10] H.-J. Butt, A technique for measuring the force between a colloidal particle in water<br />
<strong>and</strong> a bubble, J. Colloid Interface Sci. 166 (1994) 109–117.<br />
[11] W.A. Ducker, Z. Xu, J.N. Israelachvili, Measurements of hydrophobic <strong>and</strong> DLVO forces<br />
in bubble–surface interactions in aqueous solutions, Langmuir 10 (1994) 3287–3289.<br />
[12] A.V. Nguyen, G.M. Evans, J. Nalaskowski, J.D. Miller, Hydrodynamic interaction<br />
between an air bubble <strong>and</strong> a particle: atomic force microscopy measurements, Exp.<br />
Therm. Fluid Sci. 28 (2004) 387–394.<br />
[13] N.D. Wangsa-Wirawan, A. Ikai, B.K. O’Neill, A.P.J. Middelberg, Measuring the interaction<br />
forces between protein inclusion bodies <strong>and</strong> an air bubble using an atomic force microscope,<br />
Biotechnol. Prog. 17 (2001) 963–969.<br />
[14] D. Knebel, M. Sieber, H.-J. Galla, M. Amrein, Scanning force microscopy at the air–<br />
water interface of an air bubble coated with pulmonary surfactant, Biophys. J. 82 (2002)<br />
474–480.<br />
[15] N. Ishida, T. Inoue, M. Miyahara, K. Higashitani, Nano bubbles on a hydrophobic surface<br />
in water observed by tapping-mode atomic force microscopy, Langmuir 16 (2000)<br />
6377–6380.<br />
[16] P. Attard, Nanobubbles <strong>and</strong> the hydrophobic attraction, Adv. Colloid Interface Sci. 104<br />
(2003) 75–91.<br />
[17] H. Schubert, Nanobubbles, hydrophobic effect, hetrerocoagulation <strong>and</strong> hydrodynamics<br />
in flotation, Int. J. Miner. Process. 78 (2005) 11–21.<br />
[18] X.H. Zhang, X.D. Zhang, S.T. Lou, Z.X. Zhang, J.L. Sun, J. Hu, Degassing <strong>and</strong> temperature<br />
effects on the formation of nanobubbles at the mica/water interface, Langmuir<br />
20 (2004) 3813–3815.<br />
[19] N. Ishida, K. Higashitani, Interaction forces between chemically modified hydrophobic<br />
surfaces evaluated by AFM–the role of nanoscopic bubbles in the interactions,<br />
Miner. Eng. 19 (2006) 719–725.<br />
[20] V.W.A. de Villenuve, D.G.A.L. Aarts, H.N.W. Lekkerkerker, Comparing the approach of<br />
a rigid sphere <strong>and</strong> a deformable droplet towards a deformable fluid surface, Colloids<br />
Surf. A Physicochem. Eng. Asp. (282–283) (2006) 61–67.<br />
[21] P. Attard, S.J. Miklavcic, Effective spring constants of bubbles <strong>and</strong> droplets, Langmuir<br />
17 (2001) 8217–8223.<br />
[22] P. Attard, S.J. Miklavcic, Effective spring description of a bubble or a droplet interacting<br />
with a particle, J. Colloid Interface Sci. 247 (2002) 255–257.<br />
[23] X. Fan, Z. Zhang, G. Li, N.A. Rowson, Attachment of solid particles to air bubbles in<br />
surfactant-free aqueous solutions, Chem. Eng. Sci. 59 (2004) 2639–2645.<br />
[24] G. Gillies, C.A. Prestidge, P. Attard, Determination of the separation in colloid probe<br />
atomic force microscopy of deformable bodies, Langmuir 17 (2001) 7955–7956.<br />
[25] C.J. van Oss, Interfacial forces in aqueous media, first ed., Marcel Dekker Inc, 1994.<br />
[26] J.N. Israelachvili, Intermolecular <strong>and</strong> Surface Forces, second ed., Academic Press,<br />
San Diego, 1992.
104 3. QUANTIFICATION OF PARTICLE–BUBBLE INTERACTIONs<br />
[27] G.T. Barnes, I.R. Gentle, Interfacial Science, first ed., Oxford University Press, Oxford,<br />
2005.<br />
[28] R. Aveyard, J.H. Clint, Particle wettability <strong>and</strong> line tension, J. Chem. Soc. Farad. Trans.<br />
92 (1) (1996) 85–89.<br />
[29] L. Dong, D.T. Johnson, Adsorption of acicular particles at the liquid–fluid interfaces<br />
<strong>and</strong> the influence of line tension, Langmuir 21 (2005) 3838–3849.<br />
[30] E.L. Decker, Contact line structure <strong>and</strong> dynamics on surfaces with contact angle hysteresis,<br />
Langmuir 13 (1997) 6321–6332.<br />
[31] L.W. Schwartz, S. Garoff, Contact angle hysteresis on heterogeneous surfaces,<br />
Langmuir 1 (1985) 219–230.<br />
[32] J.-H. Wang, P.M. Claesson, J.L. Parker, H. Yasuda, Dynamic contact angles <strong>and</strong> contact<br />
angle hysteresis of plasma polymers, Langmuir 10 (1994) 3887–3897.<br />
[33] S.B.G. O’Brien, The meniscus near a small sphere <strong>and</strong> its relationship to line pinning<br />
of contact lines, J. Colloid Interface Sci. 183 (1996) 51–56.<br />
[34] B. Müller, M. Riedel, R. Michel, S.M. De Paul, R. Hofer, D. Heger, D. Grützmacher,<br />
Impact of nanoneter-scale roughness on contact angle hysteresis <strong>and</strong> globulin adsorption,<br />
J. Vac. Sci. Technol. B 19 (5) (2001) 1715–1720.<br />
[35] J.F. Oliver, C. Huh, S.G. Mason, Resistance to spreading of liquids by sharp edges,<br />
J. Colloid Interface Sci. 59 (3) (1977) 568–581.<br />
[36] M. Preuss, H.-J. Butt, Direct measurement of particle–bubble interactions in aqueous<br />
electrolyte: dependence on surfactant, Langmuir 14 (1998) 3164–3174.<br />
[37] M.L. Fielden, R.A. Hayes, J. Ralston, Surface <strong>and</strong> capillary forces affecting air bubble–<br />
particle interactions in aqueous electrolyte, Langmuir 12 (1996) 3721–3727.<br />
[38] A.V. Nguyen, J. Nalaskowski, J.D. Miller, The dynamic nature of contact angles as<br />
measured by atomic force microscopy, J. Colloid Interface Sci. 262 (2003) 303–306.<br />
[39] M. Preuss, H.-J. Butt, Measuring the contact angle of individual colloid particles,<br />
J. Colloid Interface Sci. 208 (1998) 468–477.<br />
[40] T. Pompe, A. Fery, S. Herminghaus, Imaging liquid structures on inhomogeneous<br />
surfaces by scanning force microscopy, Langmuir 14 (10) (1998) 2585–2588.<br />
[41] B.J. Kirby, E.F. Hasselbrink Jr., Zeta potential of microfluidic substrates: 1. Theory, experimental<br />
techniques, <strong>and</strong> effects on separations, Electrophoresis 25 (2004) 187–202.<br />
[42] P. Saulnier, J. Lachaise, G. Morel, A. Graciaa, Zeta potential of air bubbles in surfactant<br />
solutions, J. Colloid Interface Sci. 182 (1996) 395–399.<br />
[43] A. Graciaa, G. Morel, P. Saulner, J. Lachaise, R.S. Schechter, The zeta-potential of gas<br />
bubbles, J. Colloid Interface Sci. 172 (1995) 131–136.<br />
[44] D.B. Hough, L.R. White, The calculation of Hamaker constants from Lifshitz theory<br />
with applications to wetting phenomena, Adv. Colloid Interface Sci. 14 (1980) 3–41.<br />
[45] D.J. Bachman, S.J. Miklavcic, Deformation of fluid interfaces induced by electrical double<br />
layer forces <strong>and</strong> its effect on fluid–solid interactions, Langmuir 12 (1996) 4197–4204.<br />
[46] G. Gillies, M. Kapple, H.-J. Butt, Surface <strong>and</strong> capillary forces encountered by zinc<br />
sulfide microspheres in aqueous electrolyte, Langmuir 21 (2005) 5882–5886.<br />
[47] G. Gillies, K. Buscher, M. Preuss, M. Kapple, H.-J. Butt, K. Graf, Contact angles <strong>and</strong><br />
wetting behaviour of single micron sized particles, J. Phys. Condens. Matter 17 (2005)<br />
S445–S464.<br />
[48] M. Preuss, H.-J. Butt, Direct measurement of forces between particles <strong>and</strong> bubbles, Int.<br />
J. Miner. Process 56 (1999) 99–115.<br />
[49] D.J. Johnson, N. <strong>Hilal</strong>, K.E. Waters, K. Hadler, J.J. Cilliers, Measurements of interactions<br />
between particles <strong>and</strong> colloidal gas aphrons using a combined microscopic strategy,<br />
Langmuir 49 (2009) 4880–4885.<br />
[50] F. Sebba, Microfoams – an unexplained colloidal system, J. Colloid Interface Sci. 35 (4)<br />
(1971) 643–646.<br />
[51] J.J. Cilliers, D.J. Bradshaw, The flotation of fine pyrite using colloidal gas aphrons,<br />
Miner. Eng. 9 (2) (1995) 235–241.
REFERENCEs 105<br />
[52] E.A. Mansur, Y.D. Wang, Y.Y. Dai, Separation of fine particles by using colloidal gas<br />
aphrons, Chin. J. Chem. Eng. 12 (2) (2004) 286–289.<br />
[53] P.G. Chaphalkar, K.T. Valsaraj, D. Roy, A study of the size distribution <strong>and</strong> stability of<br />
colloidal gas aphrons using a particle size analyzer, Separ. Sci. Technol. 28 (6) (1993)<br />
1287–1302.<br />
[54] P.G. Chaphalkar, K.T. Valsaraj, D. Roy, Flotation using microgas dispersions for the<br />
removal of pentachlorophenol from aqueous-solutions, Separ. Sci. Technol. 29 (7)<br />
(1994) 907–921.<br />
[55] S. Ciriello, S.M. Barnett, F.J. Deluise, Removal of heavy metals from aqueous solutions<br />
using microgas dispersions, Separ. Sci. Technol. 17 (4) (1982) 521–534.<br />
[56] D. Roy, K.T. Valsaraj, S.A. Kottai, Separation of organic dyes from wastewater by<br />
using colloidal gas aphrons, Separ. Sci. Technol. 27 (5) (1992) 573–588.<br />
[57] D. Roy, K.T. Valsaraj, A. Tamayo, Comparison of soil washing using conventional surfactant<br />
solutions <strong>and</strong> colloidal gas aphrons, Separ. Sci. Technol. 27 (12) (1992) 1555–1568.<br />
[58] P. Jauregi, J. Varley, Colloidal gas aphrons: potential applications in biotechnology,<br />
TIBTECH 17 (1999) 389–395.<br />
[59] S.V. Save, V.G. Pangarkar, Harvesting of Saccharomyces cerevisiae using colloidal gas<br />
aphrons, J. Chem Technol. Biotechnol. 62 (2) (1995) 192–199.<br />
[60] E. Fuda, D. Bhatia, D.L. Pyle, P. Jauregi, Selective separation of beta-lactoglobulin from<br />
sweet whey using CGAs generated from the cationic surfactant CTAB, Biotechnol.<br />
Bioeng. 90 (5) (2005) 532–542.<br />
[61] S.V. Save, V.G. Pangarkar, S.V. Kumar, Intensification of mass-transfer in aqueous<br />
2-phase systems, Biotechnol. Bioeng. 41 (1) (1993) 72–78.<br />
[62] E. Donaldson, J. Cuy, P. Nair, B. Ratner, Poly(vinyl alcohol)-amino acid hydrogels<br />
fabricated into tissue engineering scaffolds by colloidal gas aphron technology,<br />
Macromol. Symp. 227 (2005) 115–122.<br />
[63] P. Jauregi, G.R. Mitchell, J. Varley, Colloidal gas aphrons (CGA): dispersion <strong>and</strong> structural<br />
features, AICHE J. 46 (1) (2000) 24–36.<br />
[64] L.-C. Xu, V. Vadillo-Rodriguez, B.E. Logan, Residence time, loading force, pH, <strong>and</strong><br />
ionic strength affect adhesion forces between colloids <strong>and</strong> biopolymer coated surfaces,<br />
Langmuir 21 (2005) 7491–7500.<br />
[65] D. Filip, V.I. Uricanu, M.H.G. Duits, W.G.M. Agterof, J. Mellema, Influence of bulk<br />
elasticity <strong>and</strong> interfacial tension on the deformation of gelled water-in-oil emulsion<br />
droplets: an AFM study, Langmuir 21 (2005) 115–126.<br />
[66] A.V. Nguyen, G.M. Evans, Axisymmetric approach of a solid sphere toward a nondeformable<br />
planar slip interface in the normal stagnation flow – development of<br />
global rational approximations for resistance coefficients, Int. J. Multiphas. Flow 28<br />
(2002) 1369–1380.<br />
[67] F. Benmouna, D. Johanssman, Hydrodynamic interaction of AFM cantilevers with solid<br />
walls: an investigation based on AFM noise analysis, Eur. Phys. J. E 9 (5) (2002) 435–441.<br />
[68] F. Benmouna, D. Johanssman, Hydrodynamics of particle–wall interaction in colloidal<br />
probe experiments: comparison of vertical <strong>and</strong> lateral motion, J. Phys. Condens. Matter<br />
15 (19) (2003) 3003–3012.<br />
[69] C.D.F. Honig, W.A. Ducker, No-slip hydrodynamic boundary condition for hydrophilic<br />
particles, Phys. Rev. Lett. 98 (2) (2007) 028305.<br />
[70] O.I. Vinogradova, H.-J. Butt, G.E. Yakubov, F. Feuillebois, Dynamic effects on force<br />
measurements. Viscous drag on the atomic force microscope cantilever, Rev. Sci.<br />
Instrum. 72 (2001) 2330–2339.<br />
[71] O.I. Vinogradova, G.E. Yakubov, Dynamic effects on force measurements. 2. Lubrication<br />
<strong>and</strong> the atomic force microscope, Langmuir 19 (2003) 1227–1234.<br />
[72] D.E. Aston, J.C. Berg, Thin-film hydrodynamics in fluid interface-atomic force microscopy,<br />
Ind. Eng. Chem. Res. 41 (2002) 389–396.
C H A P T E R<br />
4<br />
Investigating Membranes <strong>and</strong><br />
Membrane Processes with<br />
Atomic Force Microscopy<br />
W. <strong>Richard</strong> <strong>Bowen</strong> <strong>and</strong> <strong>Nidal</strong> <strong>Hilal</strong><br />
O U T L I N E<br />
4.1 Introduction 108<br />
4.2 The Range of Possibilities for Investigating Membranes 109<br />
4.3 Correspondence between Surface Pore Dimensions from<br />
AFM <strong>and</strong> MWCO 113<br />
4.4 Imaging in Liquid <strong>and</strong> the Determination of Surface Electrical Properties 116<br />
4.5 Effects of Surface Roughness on Interactions with Particles 120<br />
4.6 ‘Visualisation’ of the Rejection of a Colloid by a Membrane<br />
Pore <strong>and</strong> Critical Flux 123<br />
4.7 The Use of AFM in Membrane Development 124<br />
4.8 Characterisation of Metal Surfaces 126<br />
4.8.1 Effect of Electropolishing of Steel Surfaces 130<br />
4.8.2 Corrosion of Metal Surfaces 132<br />
4.8.3 Biofilms at Metal Surfaces 135<br />
4.9 Conclusions 135<br />
Acknowledgements 136<br />
List of Abbreviations 136<br />
References 136<br />
Atomic Force Microscopy in Process Engineering 107<br />
© 2009, Elsevier Ltd
108 4. INvEsTIgATINg MEMbRANEs ANd MEMbRANE PROCEssEs<br />
4.1 InTRODUCTIOn<br />
Membrane processes are one of the most significant developments in<br />
process engineering in recent times. The worldwide annual sales of membranes<br />
<strong>and</strong> membrane equipment are now worth in excess of 1 billion<br />
Euros. Membranes find widespread application in fields as diverse as water<br />
treatment, pharmaceutical processing, food processing, biotechnology,<br />
sensors <strong>and</strong> batteries. Membranes are most usually thin polymeric sheets,<br />
having pores in the range from the micrometre to sub-nanometre, that act<br />
as advanced filtration materials. This chapter is particularly concerned<br />
with pressure-driven membrane processes – microfiltration, ultrafiltration,<br />
nanofiltration <strong>and</strong> reverse osmosis. These are usually classified according<br />
to the size of materials that they separate, with ranges typically given as<br />
10.0 � 0.1 �m for microfiltration (MF), 0.1–5 nm for ultrafiltration, �1 nm for<br />
nanofiltration (NF) <strong>and</strong> �1 nm for reverse osmosis (RO).<br />
The human imagination, including the scientific imagination, is highly<br />
visual. We are most easily convinced of the existence of phenomena <strong>and</strong><br />
processes in the physical world if we can see them. The importance of seeing<br />
is deeply imbedded in language. Thus, when we finally underst<strong>and</strong><br />
some highly complex <strong>and</strong> abstract explanation, we may comment, ‘I see<br />
that now’. However, the size range of objects that the unaided human eye<br />
can directly observe is limited. We can use optical microscopes to extend<br />
the lower limit of this range down to objects of sizes comparable to the<br />
wavelength of light. Beyond this limit, we need to use other means to ‘see’.<br />
Atomic force microscopy (AFM) is one means of imaging objects of<br />
dimensions from about the wavelength of light to those below a nanometre.<br />
Thus, in the case of membranes, it is possible to visualise the membrane<br />
surface properties, such as pores <strong>and</strong> morphology, using AFM.<br />
Fortuitously, the size range of objects that may be visualised by AFM corresponds<br />
closely to the size range of surface features that determine the<br />
separation characteristics of membranes.<br />
However, the separation characteristics of membrane interfaces do not<br />
depend solely on the physical form of surface features. In liquids, surface<br />
electrical properties <strong>and</strong> the adhesion of solutes to membrane surfaces<br />
may also have profound effects on separation performance. It is thus<br />
exceedingly fortunate that an Atomic Force Microscope may also be used<br />
to determine both of these additional controlling factors. Finally, means<br />
may be devised to quantify all of these controlling factors in liquid environments<br />
that match those of process streams.<br />
The first intention of this chapter is to provide a concise review of the<br />
potential of AFM for the investigation of membranes <strong>and</strong> membrane<br />
processes, using examples from our own studies. The chapter will begin<br />
with illustrative examples that outline the range of possibilities of AFM
4.2 THE RANgE OF POssIbILITIEs FOR INvEsTIgATINg MEMbRANEs 109<br />
studies for membrane technology. Some more advanced topics will then<br />
be considered: the correspondence between surface pore dimensions<br />
from AFM <strong>and</strong> MWCO (molecular weight cut-off); imaging in liquid<br />
<strong>and</strong> the determination of surface electrical properties; effects of surface<br />
roughness on interactions with particles; ‘visualisation’ of the rejection<br />
of a colloid by a membrane pore <strong>and</strong> the use of AFM measurements in<br />
membrane development. The written text <strong>and</strong> technical details will be<br />
kept to a minimum throughout. Such details are described fully in the<br />
original publications. The intention is to let the images ‘speak’ for themselves!<br />
The second intention is to provide a short account of AFM studies<br />
of metal surfaces, especially stainless steel surfaces. Stainless steel is<br />
extensively used in the construction of membrane equipment, <strong>and</strong> the<br />
interaction of process streams with its surface can be a significant factor<br />
in the overall successful operation of the separation process. Such studies<br />
also have broader relevance due to the widespread use of stainless steel<br />
in many process industries.<br />
4.2 ThE RAngE OF POSSIbILITIES FOR<br />
InVESTIgATIng MEMbRAnES<br />
A membrane technologist acquiring an Atomic Force Microscope for<br />
the first time is best advised to begin with relatively simple measurements.<br />
A good starting point is to image some track-etch membranes in<br />
air. The pores in such membranes may also be imaged using a good optical<br />
microscope, which gives assurance about the images produced by<br />
AFM. As an example, Figure 4.1 shows an AFM image of a Cyclopore<br />
membrane of specified pore dimensions of 0.2 �m [1].<br />
µm<br />
0.20<br />
0.10<br />
0.00<br />
3<br />
µm<br />
2<br />
1<br />
0<br />
0<br />
1<br />
µm<br />
2<br />
3<br />
% of pores<br />
25<br />
20<br />
15<br />
10<br />
5<br />
0<br />
0.10 0.15 0.20 0.25 0.30 0.35<br />
Pore size, µm<br />
FIgURE 4.1 AFM image <strong>and</strong> pore size distribution of Cyclopore membrane.
110 4. INvEsTIgATINg MEMbRANEs ANd MEMbRANE PROCEssEs<br />
Shown alongside the membrane image is the derived pore size distribution.<br />
Such distributions may be readily obtained using commercial<br />
image analysis software, either automatically or manually.<br />
Once successful images of microfiltration membranes have been<br />
obtained, it is a challenge to move downwards in expected pore size or<br />
MWCO. Thus, Figure 4.2 shows a single pore in a PCI Membranes ES625<br />
ultrafiltration membrane, which has a specified MWCO of 25 000 [2].<br />
The derived pore size distribution indicates a mean pore size of 5.1 nm<br />
with a st<strong>and</strong>ard deviation of 1.1 nm. This is a pore of dimensions suitable<br />
for separating a protein molecule. It should be noted that it is not always<br />
possible to image such small pores. In particular, it is necessary for the<br />
membrane surface roughness to have characteristic dimensions less than<br />
the pore dimensions for successful imaging.<br />
An important challenge in science is ‘to boldly go’ where no man<br />
(or woman) has been before. Thus, Figure 4.3 shows an image <strong>and</strong> the<br />
Å<br />
2.00<br />
1.00<br />
0.00<br />
120.0<br />
80.0<br />
Å<br />
40.0<br />
0<br />
0<br />
40.0<br />
80.0<br />
Å<br />
120.0<br />
% of pores<br />
30<br />
20<br />
10<br />
0<br />
2 3 4 5 6 7 8 9<br />
Pore size, nm<br />
FIgURE 4.2 Single pore <strong>and</strong> pore size distribution of ES625 ultrafiltration membrane.<br />
40<br />
35<br />
30<br />
Å<br />
2.00<br />
1.00<br />
25<br />
0.00<br />
20<br />
16<br />
15<br />
12<br />
10<br />
Å 8<br />
16<br />
5<br />
4<br />
0<br />
0<br />
4<br />
8<br />
12<br />
Å<br />
0<br />
0.0 0.5 1.0 1.5<br />
Pore size, nm<br />
2.0<br />
FIgURE 4.3 Single pore <strong>and</strong> pore size distribution of XP117 nanofiltration membrane.<br />
% of pores
4.2 THE RANgE OF POssIbILITIEs FOR INvEsTIgATINg MEMbRANEs 111<br />
corresponding pore size distribution of a single pore in a nanofiltration<br />
membrane, XP117 from PCI Membranes.<br />
Caution is needed here. It is only in exceptional instances that it is possible<br />
to obtain images of such small pores. Further, some scepticism is in<br />
order. The existence of pores in microfiltration membranes may be confirmed<br />
by optical images. It does not seem unreasonable to push the limit<br />
of belief in pores down to the ultrafiltration range. But in the nanofiltration<br />
range, we need to be especially aware of the possibility of artefacts that<br />
look like pores <strong>and</strong> of seeing what we wish to visualise.<br />
It was mentioned in Section 4.1 that an Atomic Force Microscope can<br />
also quantify the other key properties controlling the membrane performance.<br />
The crucial innovation in the determination of such properties is<br />
the development of colloid probes. Such probes are formed by attaching<br />
particles of dimensions of the order of �1 �m to the end of tipless cantilevers.<br />
Such attachment may be carried out using the manipulation properties<br />
of an Atomic Force Microscope, but greater success is achieved with<br />
the use of specially designed micromanipulation equipment. An example<br />
of a colloid probe is shown in the electron microscopy image of Figure 4.4.<br />
The silica colloid probe shown is at about the lower size limit for successful<br />
micromanipulation <strong>and</strong> subsequent measurement.<br />
If such probes are manipulated to approach a single point on a membrane<br />
surface in a controlled manner in an electrolyte solution, it is possible<br />
to directly quantify the electrical double layer interactions between probe<br />
<strong>and</strong> membrane, also shown in Figure 4.4, for two solutions of differing ionic<br />
strengths. Such electrical interactions are very important in determining the<br />
rejection of colloids <strong>and</strong> biological macromolecules during ultrafiltration.<br />
The adhesion of process stream components to membranes also has an<br />
important influence on membrane performance. Ideally, such adhesion<br />
(F/R) / (mN/m)<br />
20<br />
15<br />
10<br />
5<br />
10-4 10<br />
M NaCl, pH8<br />
-1 M NaCl, pH8<br />
0<br />
0 5 10 15 20 25 30 35 40<br />
Distance / nm<br />
FIgURE 4.4 SEM image of a 0.75 �m silica sphere attached to a tipless AFM cantilever<br />
<strong>and</strong> force distance curves for the approach of the probe to a Cyclopore microfiltration membrane<br />
in NaCl solutions at pH 8. (F is force <strong>and</strong> R is sphere radius.)
112 4. INvEsTIgATINg MEMbRANEs ANd MEMbRANE PROCEssEs<br />
–3<br />
1000 1200 1400 1600 1800 2000<br />
Piezo displacement [nm]<br />
should be avoided. As an example, Figure 4.5 shows a polystyrene colloid<br />
probe <strong>and</strong> data for the retraction of such a probe after it has been<br />
pushed into contact with membrane surfaces [3].<br />
The depths of the depressions of the curves in Figure 4.5 give direct<br />
quantification of the adhesion of the probes to the surfaces. The ES404<br />
membrane was an existing commercial membrane <strong>and</strong> the XP117 membrane<br />
is a development membrane designed to have lower fouling properties<br />
(both membranes are PCI Membranes). The development membrane<br />
had significantly lower adhesion <strong>and</strong> hence significantly lower fouling<br />
potential than the existing membrane.<br />
Ultrafiltration membranes are used particularly for the processing of<br />
solutions of biological macromolecules. Such molecules are too small to<br />
immobilise singly on a cantilever. However, by adsorbing such molecules<br />
onto colloid probes, it is possible to measure their adhesion to membrane<br />
surfaces. Such a protein (BSA – bovine serum albumin)-modified probe<br />
<strong>and</strong> the corresponding adhesion data for two membranes is shown in<br />
Figure 4.6 [4].<br />
The data shows that the protein-modified probe has significantly<br />
lower adhesion to the development membrane, XP117, than to the existing<br />
commercial membrane, ES404, <strong>and</strong> hence the modified membrane<br />
has significantly lower fouling potential in the processing of such protein<br />
solutions.<br />
Membranes are frequently used to process biological cell dispersions.<br />
In order to elucidate such processes, it has been proved possible<br />
to immobilise single cells at tipless AFM cantilevers, creating cell probes,<br />
whilst maintaining the viability of such a cell, Figure 4.7 [5].<br />
Such cell probes allow the direct measurement of the adhesion of biological<br />
cells to membranes. Interpretation of the data for such cell probes<br />
F/R [mN/m]<br />
5<br />
4<br />
3<br />
2<br />
1<br />
0<br />
–1<br />
–2<br />
ES 404<br />
XP 117<br />
FIgURE 4.5 SEM image of a 11 �m polystyrene sphere attached to a tipless AFM cantilever<br />
<strong>and</strong> force versus piezo displacement plot (retraction) for two membranes in 10 �2 M<br />
NaCl solution at pH 8.0 (ES404, conventional; XP117, modified).<br />
D<br />
C<br />
B<br />
A'<br />
A
4.3 CORREsPONdENCE bETwEEN sURFACE PORE dIMENsIONs FROM AFM ANd MwCO 113<br />
F/R [mN/m]<br />
–30<br />
0 200 400 600 800<br />
Piezo displacement [nm]<br />
is more difficult than for colloid or modified colloid probes as the cell can<br />
distort during the measurements. However, the data show that the XP117<br />
development membrane has lower adhesion, <strong>and</strong> hence lower fouling<br />
propensity, also for yeast cells.<br />
4.3 CORRESPOnDEnCE bETWEEn SURFACE PORE<br />
DIMEnSIOnS FROM AFM AnD MWCO<br />
From a practical point of view, the choice of ultrafiltration membranes<br />
for a particular process is usually defined in terms of MWCO, defined<br />
such that solutes of such molecular weight would be 90% rejected by the<br />
30<br />
20<br />
10<br />
0<br />
–10<br />
–20<br />
Modified membrane (2)<br />
Conventional membrane (1)<br />
FIgURE 4.6 SEM image of a BSA-coated silica sphere attached to a tipless AFM cantilever<br />
<strong>and</strong> force versus piezo displacement plot (retraction) for two membranes in 10 �2 M<br />
NaCl solution at pH 5.0 (conventional, ES404; modified, XP117).<br />
Force [nN]<br />
20<br />
15<br />
10<br />
5<br />
0<br />
–5<br />
C<br />
B<br />
Conventional membrane (1)<br />
Modified membrane (2)<br />
–10<br />
0 100 200 300 400 500<br />
Piezo displacement [nm]<br />
FIgURE 4.7 SEM image of a yeast cell (Saccharomyces cerevisiae) attached to a tipless<br />
AFM cantilever – a cell probe – <strong>and</strong> force versus piezo displacement plot (retraction) for<br />
two membranes in 10 �2 M NaCl solution at pH 5.0 (conventional, ES404; modified, XP117).
114 4. INvEsTIgATINg MEMbRANEs ANd MEMbRANE PROCEssEs<br />
membrane. Care is required in the use of MWCO as rejection can depend<br />
on other factors such as charge <strong>and</strong> molecular shape, but it remains an<br />
extensively used parameter. It is possible to estimate the mean pore diameter<br />
of membranes from rejection data such as MWCO through the use<br />
of an appropriate mathematical expression. As MWCO is a historically<br />
important measure <strong>and</strong> AFM a relatively new technique, it is pertinent to<br />
investigate the correspondence between the data respectively obtained.<br />
Such an investigation has been carried out for a series of membranes<br />
with MWCO in the range 1000–10 000 (Desal-G, Osmonics) [5]. Images<br />
of membranes obtained in air at the extremes of this range, GE (1000)<br />
<strong>and</strong> GN (10 000), are shown in Figure 4.8. The higher MWCO membrane<br />
is noticeably rougher. Indeed, there is a significant increase in surface<br />
roughness throughout the range, as shown by the data in Table 4.1.<br />
From such images, it is possible to determine the surface pore diameter<br />
distributions of the membranes. Data is given in Table 4.2, together with<br />
the mean pore diameters from MWCO, using three differing mathematical<br />
Å<br />
268<br />
134<br />
0<br />
3<br />
2<br />
300<br />
µm<br />
Å<br />
3.00<br />
2.00<br />
1.00<br />
0.00<br />
1<br />
200<br />
Å<br />
GE membrane<br />
100<br />
0 0<br />
1<br />
µm<br />
2<br />
3<br />
3<br />
2<br />
µm<br />
1<br />
GN membrane<br />
0 0<br />
GE membrane GN membrane<br />
Å<br />
2.00<br />
0 0<br />
100<br />
Å<br />
200<br />
300<br />
300<br />
200<br />
100<br />
0 0<br />
FIgURE 4.8 AFM images of two Desal G membranes, GE (MWCO 1000) <strong>and</strong> GN<br />
(MWCO 10 000), at low resolution (above) <strong>and</strong> high resolution (below).<br />
1.00<br />
0.00<br />
Å<br />
1<br />
100<br />
µm<br />
2<br />
Å<br />
200<br />
3<br />
300
4.3 CORREsPONdENCE bETwEEN sURFACE PORE dIMENsIONs FROM AFM ANd MwCO 115<br />
expressions. Depending on the expression used, the ratio of the mean pore<br />
diameters obtained from MWCO data <strong>and</strong> AFM measurements was in the<br />
range 1.04–2.42. Agreement between the diameters obtained in the two ways<br />
was best for the membranes with the lowest MWCO values. The two measurements<br />
are probing the membrane properties in different ways, but the<br />
overall correlation is encouraging.<br />
However, AFM has the important advantage of providing a measure<br />
of pore size distribution. This is indicated in Table 4.2 by the<br />
st<strong>and</strong>ard deviations. A full distribution for one of the membranes is given<br />
in Figure 4.9.<br />
Knowledge of such distributions is very important if the membrane is<br />
to be selected for a fractionation process, where a narrow distribution is<br />
highly desirable. Further, theoretical modelling to predict membrane process<br />
performance often assumes a log-normal distribution of pore sizes.<br />
Figure 4.9 shows that this is a reasonable assumption in this case.<br />
TAbLE 4.1 surface Characteristics of desal g Membranes Over Areas of<br />
3 �m � 3 �m (R p−v, Peak to valley distance; Rms, Root-Mean-square).<br />
Membrane MWCO R p–v (nm) Rms roughness (nm)<br />
GE 1000 26.6 3.6<br />
GH 2500 58.7 8.4<br />
GK 3500 53.9 8.4<br />
GM 8000 63.7 9.2<br />
GN 10 000 88.1 11.7<br />
TAbLE 4.2 Pore diameters for desal g-series Membranes Obtained from AFM<br />
Measurements <strong>and</strong> MwCO.<br />
AFM mean pore<br />
diameter Mean pore diameters from MWCO (nm)<br />
Membrane (nm) * Ex. 1 Ex. 2 Ex. 3<br />
GE 1.8 (�0.3) 1.9 2.4 2.0<br />
GH 2.2 (�0.5) 2.6 3.8 3.2<br />
GK 2.5 (�0.5) 3.0 4.4 4.2<br />
GM 2.8 (�0.7) 4.4 6.4 6.8<br />
GN 3.1 (�0.9) 4.8 7.2 7.6<br />
* Numbers in brackets are st<strong>and</strong>ard deviations.
116 4. INvEsTIgATINg MEMbRANEs ANd MEMbRANE PROCEssEs<br />
Fraction of pores counted<br />
0.16<br />
0.14<br />
0.12<br />
0.10<br />
0.08<br />
0.06<br />
0.04<br />
0.02<br />
AFM distribution<br />
Theoretical distribution<br />
0.00<br />
0.00<br />
2.0 2.5 3.0 3.5 4.0 4.5 5.0<br />
Pore diameter/nm<br />
FIgURE 4.9 AFM pore size distribution for Desal GM membrane <strong>and</strong> theoretical lognormal<br />
distribution.<br />
4.4 IMAgIng In LIqUID AnD ThE DETERMInATIOn OF<br />
SURFACE ELECTRICAL PROPERTIES<br />
For membranes that are used in liquid systems, it is very useful to<br />
have the possibility of imaging in a solution corresponding to the processing<br />
conditions of pH <strong>and</strong> ionic content. An Atomic Force Microscope<br />
gives great control of imaging protocols, allowing investigation of the<br />
procedures that give the best membrane images, in particular the imaging<br />
force. Figure 4.10 shows the force of interaction between an AFM tip<br />
<strong>and</strong> a Cyclopore membrane of specified pore diameter 0.1 �m in solutions<br />
at constant pH but varying ionic strength [6].<br />
It may be seen that the range of the interaction increases greatly as the<br />
ionic strength decreases in accordance with electrical double layer theory. It<br />
is then possible to image the membrane using a force at any point on the<br />
curves in Figure 4.10. Two series of such images, one at various forces at<br />
constant ionic strength <strong>and</strong> the other at approximately constant force at various<br />
ionic strengths, are shown in Figure 4.11. It may be seen that the quality<br />
of the images improves with increasing imaging force <strong>and</strong> with increasing<br />
ionic strength. The derived pore diameter distributions also vary with the<br />
imaging conditions, as shown in the corresponding graphs in Figure 4.11 [7].<br />
The reason for this variation may be understood from Figure 4.12,<br />
which shows calculated isopotential lines for a 10 �1 M solution <strong>and</strong> a<br />
10 �4 M solution, respectively.<br />
0.16<br />
0.14<br />
0.12<br />
0.10<br />
0.08<br />
0.06<br />
0.04<br />
0.02<br />
Probability density function fR(r)
4.4 IMAgINg IN LIqUId ANd THE dETERMINATION OF sURFACE ELECTRICAL 117<br />
Force (nN)<br />
50<br />
40<br />
30<br />
20<br />
10<br />
NaCl 10 –1 M<br />
NaCl 10 –2 M<br />
NaCl 10 –3 M<br />
NaCl 10 –4 M<br />
0<br />
0 5 10 15 20 25 30 35<br />
Distance (nm)<br />
FIgURE 4.10 Force versus distance curves for the approach of an AFM tip to a<br />
Cyclopore microfiltration membrane in NaCl solutions of various ionic strengths at pH 6.5.<br />
Roughly understood, the imaging tip will follow an isopotential line<br />
during the imaging process. At the high ionic strength, all of the calculated<br />
lines lie close to the membrane surface (shaded, spinning the image<br />
360° out of the plane of the paper would generate a pore). At the lower<br />
ionic strength, some of the isopotential lines lie far from the surface, so<br />
a clear, veracious pore image would not be obtained when such a line is<br />
followed, as would be the case at low imaging forces. In conclusion, the<br />
best imaging conditions in ionic solutions are at high imaging force <strong>and</strong>,<br />
if possible <strong>and</strong> appropriate, at high ionic strength.<br />
Though most membrane technologists now recognise the important<br />
contribution of membrane surface electrical properties in defining the<br />
separation characteristics, there remains confusion as to how best to<br />
describe <strong>and</strong> quantify such interactions. There is an unfortunate tendency<br />
in the applied membrane literature to use the terms ‘charge’ <strong>and</strong> ‘potential’<br />
loosely, almost interchangeably. There is also a regrettable lack of<br />
precision in the interpretation of membrane streaming potentials, which<br />
are the basic data most commonly used to quantify membrane surface<br />
electrical properties. The interpretation of streaming potential data can<br />
be complex even for perfectly smooth <strong>and</strong> chemically uniform planar<br />
surfaces. Most membrane surfaces have roughness comparable to electrical<br />
double layer dimensions, so along-the-surface-membrane streaming<br />
potential data give only some average property. Through-the-membrane<br />
streaming potential data may only be usefully interpreted in comparison
118 4. INvEsTIgATINg MEMbRANEs ANd MEMbRANE PROCEssEs<br />
Imaging force = 4.35 nN<br />
NaCl 10 -1<br />
2<br />
µm<br />
1<br />
%<br />
A<br />
0<br />
2<br />
1.5<br />
µm 1<br />
0.5<br />
0<br />
50<br />
40<br />
30<br />
20<br />
10<br />
0<br />
Imaging force = 33 nN<br />
NaCl 10 -4<br />
0<br />
0.5<br />
1<br />
µm<br />
IF = 4.4 nN<br />
10 -1 M<br />
0<br />
0.05 0.10<br />
Pore size, µm<br />
0.15<br />
2<br />
1.5<br />
1 µm<br />
2<br />
Imaging force = 12.9 nN<br />
NaCl 10 -1<br />
2<br />
µm<br />
A B C<br />
Imaging force = 32.1 nN<br />
NaCl 10 -3<br />
2<br />
1.5<br />
µm 1<br />
0.5<br />
0<br />
0<br />
0.5<br />
1.5<br />
1<br />
µm<br />
D E F<br />
%<br />
50<br />
40<br />
30<br />
20<br />
10<br />
IF = 33.0 nN<br />
10 -4 M<br />
0<br />
0.05 0.10 0.15 0.20<br />
Pore size, µm<br />
%<br />
%<br />
50<br />
40<br />
30<br />
20<br />
10<br />
1<br />
0<br />
0<br />
with numerical double layer calculations, as pore diameters are often less<br />
than the characteristic electrical double layer dimensions.<br />
Fortunately, AFM, in conjunction with the colloid probe technique,<br />
offers an alternative means of membrane surface electrical properties<br />
characterisation. If a colloid probe is approached towards a surface, it<br />
is possible to quantify the force of interaction. Typical data is shown in<br />
Figure 4.13 for a Desal DK membrane, which is one of the least rough<br />
membranes [7].<br />
1<br />
µm<br />
IF = 12.9 nN<br />
10 -1 M<br />
Pore size, µm<br />
B C<br />
IF = 33.1 nN<br />
10 -3 M<br />
0<br />
0.05 0.10 0.15<br />
Pore size, µm<br />
D E F<br />
50<br />
40<br />
30<br />
20<br />
10<br />
0<br />
0.05 0.10 0.15<br />
2<br />
2<br />
%<br />
50<br />
40<br />
30<br />
20<br />
10<br />
Imaging force = 33.8 nN<br />
NaCl 10 -1<br />
2<br />
µm<br />
1<br />
0<br />
Imaging force = 33.2 nN<br />
NaCl 10 -2<br />
2<br />
1.5<br />
µm 1<br />
0.5<br />
0<br />
0<br />
0<br />
IF = 33.8 nN<br />
10 -1 M<br />
0.5<br />
0<br />
0.05 0.10 0.15<br />
Pore size, µm<br />
IF = 33.2 nN<br />
10 -2 M<br />
0<br />
0.05 0.10 0.15<br />
Pore size, µm<br />
1<br />
µm<br />
1<br />
1.5<br />
µm<br />
FIgURE 4.11 AFM images of a Cyclopore membrane in electrolyte solutions. (a)–(c) at<br />
various forces in 10 �1 M NaCl solution; (c), (d)–(f) at approximately constant force at various<br />
ionic strengths. All data at pH 6.5 (IF-imaging force).<br />
%<br />
50<br />
40<br />
30<br />
20<br />
10<br />
2<br />
2
4.4 IMAgINg IN LIqUId ANd THE dETERMINATION OF sURFACE ELECTRICAL 119<br />
F<br />
A<br />
A B<br />
0.3<br />
0.4<br />
C D 0.7<br />
0.4<br />
0.1<br />
0.5<br />
(a)<br />
C D<br />
F E<br />
FIgURE 4.12 Calculated isopotential lines at the entrance to a membrane pore of diameter<br />
0.1 �m. (a) 10 �1 M solution, (b) 10 �2 M solution.<br />
F/R/mN/m<br />
4<br />
2<br />
0<br />
0 20 40<br />
B<br />
Distance/nm<br />
(b)<br />
0.6<br />
0.7<br />
0.9<br />
0.8<br />
AFM data<br />
Best fit constant potential<br />
Best fit constant charge<br />
FIgURE 4.13 Best-fit solutions of theoretical force distance curves to AFM experimental<br />
force distance data for a Desal membrane in 10 �3 M NaCl.<br />
E
120 4. INvEsTIgATINg MEMbRANEs ANd MEMbRANE PROCEssEs<br />
Moreover, if the surface potential or surface charge of the colloid has<br />
been determined <strong>and</strong> the solution is of defined ionic content, it is possible<br />
to calculate the potential or charge of the surface under investigation<br />
by matching the experimentally obtained curves to theoretical calculations<br />
on the basis of electrical double layer theory. In the example shown,<br />
the best-fit membrane surface charge was �0.00114 C m �2 <strong>and</strong> the bestfit<br />
membrane surface potential was �64 mV. Furthermore, an important<br />
advantage of the colloid probe technique is that it allows exploration<br />
of variations in surface electrical interactions at different points on the<br />
membrane surface, as the following section shows.<br />
4.5 EFFECTS OF SURFACE ROUghnESS On<br />
InTERACTIOnS WITh PARTICLES<br />
Surfaces are usually imaged using sharp tips in order to produce<br />
images with the highest possible definition. However, from a processing<br />
viewpoint, an important factor is how process stream components such as<br />
solutes <strong>and</strong> colloidal particles interact with the surface. If a surface has feature<br />
variations that have dimensions comparable with those of the process<br />
stream components, then such interactions may show variations at different<br />
locations on the surface. An effective way of gauging such effects is to<br />
image the surface with an appropriate stream component, e.g. a particle<br />
that is immobilised to form a colloid probe. Figure 4.14 shows results for<br />
a reverse osmosis membrane (AFC99, PCI Membranes) imaged in saline<br />
solution with first a sharp tip <strong>and</strong> then a 4.2 �m silica sphere [8].<br />
µm<br />
0.4<br />
0.2<br />
4<br />
0<br />
3<br />
P–v = 560 nm<br />
Rms = 66nm<br />
2<br />
µm<br />
1 1<br />
0 0<br />
2<br />
µm<br />
A B<br />
3<br />
4<br />
µm<br />
0.16<br />
0.12<br />
0.2<br />
0.04<br />
4<br />
0<br />
3<br />
2<br />
µm<br />
P–v = 186 nm<br />
Rms = 21nm<br />
1 1<br />
0 0<br />
FIgURE 4.14 AFC99 membrane imaged with a tip (A) <strong>and</strong> with a 4.2 �m colloid probe<br />
(B) (P–v, peak to valley; Rms, root mean square).<br />
2<br />
µm<br />
3<br />
4
4.5 EFFECTs OF sURFACE ROUgHNEss ON INTERACTIONs wITH PARTICLEs 121<br />
The membrane surface features are still apparent but less well defined<br />
in the colloid probe image. This is a rather rough membrane showing<br />
clear peaks <strong>and</strong> valleys. Following imaging, it is possible to position<br />
the colloid probe at any point on the surface to quantify colloid–surface<br />
interactions. Figure 4.15 shows long-range electrostatic interactions quantified<br />
at peaks <strong>and</strong> valleys respectively, <strong>and</strong> some analysis of the data is<br />
presented in Table 4.3.<br />
Both the magnitude <strong>and</strong> range of the electrical double layer interactions<br />
on the peaks are greatly reduced compared to that in the valleys. In<br />
both cases, the range is very different from that at a planar surface, both<br />
experimental – data for mica is shown for comparison – or theoretical,<br />
the Debye length. It is also possible to quantify the adhesive interaction<br />
between colloid probe <strong>and</strong> membrane surface at different locations, as<br />
shown in Figure 4.16 <strong>and</strong> Table 4.4.<br />
F/R/mN/m<br />
10<br />
1<br />
AFC99, peaks<br />
10 –1 M<br />
10 –2 M<br />
10 –3 M<br />
0.1<br />
0 10 20<br />
Distance/nm<br />
30 40<br />
F/R/mN/m<br />
10<br />
1<br />
AFC99, valleys<br />
10 –1 M<br />
10 –2 M<br />
10 –3 M<br />
0.1<br />
0 10 20<br />
Distance/nm<br />
30 40<br />
FIgURE 4.15 Long-range electrostatic interactions at peaks <strong>and</strong> valleys for an AFC99<br />
membrane at various ionic strengths in NaCl solutions (F, force; R, sphere radius).<br />
TAbLE 4.3 Effective Experimental decay Lengths for Approach Curves between<br />
silica Colloid Probe <strong>and</strong> AFC99 Membrane <strong>and</strong> Mica, <strong>and</strong> debye Length.<br />
Effective decay lengths (nm)<br />
[NaCl] (M) Membrane peak Membrane valley Mica Debye length (nm)<br />
10 �3 7.7�1.4 12.1�1.5 9.2�0.5 9.6<br />
10 �2 3.2�0.6 6.8�1.2 3.5�0.3 3.1
122 4. INvEsTIgATINg MEMbRANEs ANd MEMbRANE PROCEssEs<br />
The adhesion of the colloid probe is markedly lower at the peaks on<br />
the membrane surface than in the valleys, with the difference increasing<br />
with decreasing salt concentration, <strong>and</strong> reaching a factor of more than 20<br />
in 10 �3 M solution. The wide variation in interactions shows that theoretical<br />
descriptions of membrane fouling need to take surface morphology<br />
into explicit account. Further, the results show that the selection of membranes<br />
for the filtration of specific process streams would benefit from an<br />
assessment of the size of likely foulants <strong>and</strong> the membrane roughness,<br />
especially the periodicity of the roughness. Fouling could be minimised<br />
F/R / mN/m<br />
6<br />
4<br />
2<br />
0<br />
–2<br />
–4<br />
–6<br />
Peak<br />
Valley<br />
–8<br />
0 50 100 150<br />
Separation distance / nm<br />
FIgURE 4.16 Adhesion of a silica colloid probe at a peak <strong>and</strong> a valley on an AFC99<br />
membrane in 10 �3 M NaCl solution (pull-off force).<br />
TAbLE 4.4 Normalised Adhesion Forces for silica Colloid Probes <strong>and</strong> AFC99<br />
Membrane.<br />
Surface type [NaCl] (M) F/R (mN/m)<br />
Membrane peak 10 �3 �0.3 (�0.17)<br />
10 �2 �1.4 (�0.43)<br />
10 �1 �2.3 (�0.48)<br />
Membrane valley 10 �3 �6.5 (�2.2)<br />
10 �2 �8.1 (�3.5)
4.6 ‘vIsUALIsATION’ OF THE REjECTION OF A COLLOId by A MEMbRANE PORE 123<br />
by using membranes with roughness such that only adhesion at peaks is<br />
possible – where the effects of cross-flow are also maximum.<br />
4.6 ‘VISUALISATIOn’ OF ThE REjECTIOn OF A COLLOID<br />
by A MEMbRAnE PORE AnD CRITICAL FLUx<br />
One of the most useful practical operating concepts for membrane processes<br />
is that of a critical filtration flux or critical operating pressure. These<br />
critical parameters are such that below such critical values, rejection will<br />
occur <strong>and</strong> fouling will be minimum <strong>and</strong> above these critical values, transmission<br />
<strong>and</strong> fouling may take place. For colloidal particles, the critical<br />
values may arise as a balance between the hydrodynamic force driving<br />
the solutes towards a membrane pore <strong>and</strong> an electrostatic (electrical<br />
double layer) force opposing this motion.<br />
Science fiction writers have imagined tiny probes travelling, for example,<br />
through the human body, which would allow us to visualise hidden microscopic<br />
phenomena. Such probes remain figments of the imagination.<br />
However, a colloid probe moving along a membrane surface can allow us to<br />
visualise how a colloidal particle would ‘see’ such a surface, Figure 4.17 [9].<br />
Figure 4.17 shows how a 0.75 �m silica colloid probe ‘sees’ the pores in<br />
a 1.0 �m Cyclopore membrane in solutions of two ionic strengths when<br />
imaged in each case with an applied force of 4.6 mN m –1 . It is only at the<br />
highest ionic strength that the pores are clearly apparent, for at the lowest<br />
ionic strength, such a force is experienced too far from the membrane surface<br />
for the electric field to still show sufficient evidence of membrane<br />
porosity. Cases where the process stream component dimensions are close<br />
to those of the membrane pores should preferably be avoided as rapid<br />
pore blocking may occur if the critical flux or pressure is exceeded. Such<br />
6<br />
4<br />
µm<br />
6<br />
4<br />
µm<br />
2<br />
6<br />
2<br />
0<br />
2<br />
4<br />
µm<br />
0<br />
0<br />
FIgURE 4.17 Imaging a 1.0 �m Cyclopore membrane with a 0.75 �m colloid probe, in<br />
0.1 M NaCl, pH 8 (left); in 0.0001 M NaCl, pH 8 (right). Normalised imaging force 4.6mN m –1 .<br />
0<br />
2<br />
4<br />
µm<br />
6
124 4. INvEsTIgATINg MEMbRANEs ANd MEMbRANE PROCEssEs<br />
Flux / ms –1<br />
0.0006<br />
0.0005<br />
0.0004<br />
0.0003<br />
0.0002<br />
0.0001<br />
0.0000<br />
0 5 10 15 20 25<br />
Time / min<br />
250 kPa<br />
50 kPa<br />
FIgURE 4.18 Filtration fluxes as a function of time for filtration of particles very close<br />
in dimensions to the pore dimensions. Silica particles, 0.001 M NaCl, pH 6.0, mean particle<br />
diameter 86 nm, mean pore diameter 85 nm, calculated critical pressure � 130 kPa.<br />
conditions may, however, be theoretically predicted. Thus, Figure 4.18 shows<br />
the time dependence of flux for such a case where the critical pressure is calculated<br />
to be 130 kPa. Above this pressure, there is a rapid decrease in flux<br />
with time, but below this pressure, only a very gradual flux decrease occurs.<br />
4.7 ThE USE OF AFM In MEMbRAnE DEVELOPMEnT<br />
The ability of AFM to guide the development of improved membranes<br />
has been shown in the development of polysulphone-sulphonated<br />
poly(ether ether) ketone (PSU/SPEEK) blend membranes. The aim of the<br />
work was to develop highly charged membranes with correspondingly<br />
low adhesion characteristics <strong>and</strong> high critical fluxes [10]. The SPEEK provides<br />
negative charges, as shown by the structure in Figure 4.19.<br />
One great benefit of SPEEK is that it acts as a pore formation–promoting<br />
agent. This gives membranes high porosity <strong>and</strong> high fluxes, as shown by the<br />
increase in water flux, with increasing SPEEK content shown in Figure 4.20.<br />
The increase in porosity may be directly confirmed by high-resolution<br />
images of an unmodified PSU membrane <strong>and</strong> a PSU/SPEEK membrane<br />
with 5% of the charged polymer, Figure 4.21.
4.7 THE UsE OF AFM IN MEMbRANE dEvELOPMENT 125<br />
CH 3<br />
O C O S<br />
CH3<br />
PSU<br />
O O C<br />
SO 3H<br />
SPEEK<br />
FIgURE 4.19 Structures of polysulphone <strong>and</strong> sulphonated poly (ether ether) ketone.<br />
Jw / ms –1<br />
1.2e−4<br />
1.0e−4<br />
8.0e−5<br />
6.0e−5<br />
4.0e−5<br />
2.0e−5<br />
P-O<br />
P-P<br />
S0.5-20<br />
S2-20<br />
S5-20<br />
ΔP / kN m –2<br />
0.0e+0<br />
0 100 200 300 400 500 600<br />
FIgURE 4.20 Water flux as a function of applied pressure for five membranes of various<br />
percentages of SPEEK. P-O <strong>and</strong> P-P: 0%; S0.5-20: 0.5%; S2-20: 2.0%; S5-20: 5%.<br />
The adhesion of a range of colloid probes, both inorganic <strong>and</strong> biological,<br />
is greatly reduced at the PSU/SPEEK membranes, as shown by the<br />
data in Table 4.5.<br />
Such low adhesion forces show that the membranes are well suited to<br />
many types of separation. The removal of humic acid has been investigated<br />
as an example of a challenging separation. The blend membranes<br />
O<br />
O<br />
O<br />
n<br />
n
126 4. INvEsTIgATINg MEMbRANEs ANd MEMbRANE PROCEssEs<br />
300<br />
200<br />
Å<br />
100<br />
0 0<br />
100<br />
Å<br />
200<br />
300<br />
gave very good separation with little fouling. By quantification of the<br />
fraction of humic acids deposited during filtration, it was possible to<br />
show that both planar (S5-20) <strong>and</strong> tubular (T5-20) versions of the membranes<br />
showed high critical flux values, whereas membranes already in<br />
commercial use for such separations (CA202 <strong>and</strong> ES404, PCI Membranes)<br />
did not show such a favourable property, Figure 4.22 [11].<br />
To allow for the different membrane geometries, the data in Figure 4.22<br />
is expressed as critical Peclet numbers (J/k, where J is the flux <strong>and</strong> k is<br />
the mass transfer coefficient). The critical Peclet numbers were 4.6 <strong>and</strong><br />
6.4 for the planar <strong>and</strong> tubular membranes respectively. Further details of<br />
the development of PSU/SPEEK membranes are given in Chapter 5.<br />
4.8 ChARACTERISATIOn OF METAL SURFACES<br />
The surface properties <strong>and</strong> morphology of process equipment surfaces<br />
are of fundamental importance in the fulfilment of their purpose <strong>and</strong><br />
300<br />
200<br />
Å<br />
100<br />
0<br />
100<br />
200<br />
Å<br />
FIgURE 4.21 High-resolution images of unmodified membrane (left) <strong>and</strong> membrane<br />
modified with 5% SPEEK (right, S5-20).<br />
TAbLE 4.5 Normalised Adhesion Forces for a Range of Colloid Probes at a PsU/<br />
sPEEK Membrane (s5-20).<br />
Materials Mean F/R (mN m �1 )<br />
SPEEK/PSU PSU<br />
Silica 0.84�0.35 6.5�0.8<br />
Cellulose 0.51�0.23 2.2�0.32<br />
Latex 2.1�0.52 14.3�0.8<br />
BSA 1.4�0.64 12.3�1.4<br />
Yeast 3.2�0.85 9.5�1.5<br />
300
Dfractional<br />
0.18<br />
0.16<br />
0.14<br />
0.12<br />
0.10<br />
0.08<br />
0.06<br />
0.04<br />
0.02<br />
4.8 CHARACTERIsATION OF METAL sURFACEs 127<br />
CA202, ES404<br />
S5-20<br />
T5-20<br />
0.00<br />
0 5 10 15<br />
JHA / k<br />
20 25 30<br />
FIgURE 4.22 Fractional deposition of humic acids at different operating fluxes. The<br />
intercepts on the x-axis indicate the critical fluxes.<br />
applications in (bio) process industries. Many environments to which<br />
they may become exposed are potentially corrosive, <strong>and</strong> thus more<br />
corrosion-resistant surfaces may have greater operational lifetimes. In<br />
addition, surfaces may become fouled, leading to a reduction in operating<br />
efficiency, especially for surfaces involved in the desalination <strong>and</strong><br />
water treatment industries. This section will concentrate on the use of<br />
AFM to characterise the surface features <strong>and</strong> morphology of metal surfaces<br />
of interest in process engineering applications.<br />
One of the most commonly used materials in process equipment is<br />
stainless steel in various forms. In Figure 4.23, by way of example, are<br />
shown three 3-dimensional 50 � 50 �m scans of different steel surfaces of<br />
difference roughness characteristics, imaged in air [12].<br />
These are: (a) a ‘highly polished’ stainless steel surface with an optically<br />
fine finish, (b) the surface of a stainless steel sample disk, such as<br />
that commonly used for affixing AFM samples <strong>and</strong> (c) a sample stub<br />
similar to that in (b), which has undergone significant pitting owing to<br />
corrosion. Surface (b) contains ridges caused by the mechanical polishing<br />
of the surface, which are not visible in (a). For the corroded sample,<br />
there is clearly a much greater variation in the height of surface<br />
features <strong>and</strong> a greater number of asperities. This can be seen from the<br />
roughness parameters obtained from the height data contained in these
128 4. INvEsTIgATINg MEMbRANEs ANd MEMbRANE PROCEssEs<br />
50 µm<br />
A<br />
50 µm<br />
B<br />
50 µm<br />
C<br />
25 µm<br />
25 µm<br />
25 µm<br />
0 µm<br />
0 µm<br />
0 µm<br />
0 µm<br />
0 µm<br />
0 µm<br />
25 µm<br />
25 µm<br />
25 µm<br />
726.59 nm<br />
0 nm<br />
50 µm<br />
1.47 µm<br />
0.73 µm<br />
0 µm<br />
50 µm<br />
3.56 µm<br />
1.78 µm<br />
0 µm<br />
50 µm<br />
FIgURE 4.23 3D topographic images of surfaces with three different grades of roughness.
4.8 CHARACTERIsATION OF METAL sURFACEs 129<br />
TAbLE 4.6 surface Characteristics Obtained from AFM Images.<br />
Surface Area Ra (nm) RMS (nm) Average height (nm) Maximum range (nm)<br />
(a) 38.3 49.5 361.1 726.6<br />
(b) 124.3 170.1 785.4 1407.5<br />
FIgURE 4.24 SEM image of a CaCO 3 crystal mounted as a probe on the apex of an<br />
AFM cantilever.<br />
images. Table 4.6 shows surface characteristics obtained from such<br />
AFM images.<br />
The differences in the roughness average (Ra) <strong>and</strong> root mean square<br />
roughness (RMS) show that the statistical variation in height of the<br />
individual points in each image becomes greater from sample (a) to (c),<br />
whilst the average height <strong>and</strong> maximum range show that the magnitude<br />
of the variations in height, <strong>and</strong> hence size, of the surface features is much<br />
greater from sample (a) to (c).<br />
Adhesion measurements were taken with a CaCO 3 crystal, Figure 4.24,<br />
at five different points on the sample surface, with a total of approximately<br />
100 measurements per sample, <strong>and</strong> mean adhesion values were calculated.
130 4. INvEsTIgATINg MEMbRANEs ANd MEMbRANE PROCEssEs<br />
Figure 4.25 shows a summary of the adhesion data measured between<br />
the probe <strong>and</strong> each of the stainless steel surfaces of interest.<br />
As can be seen, sample (c) has the lowest adhesion value measured,<br />
with the probe detaching at a mean force of 65.5 nN. Samples (b) <strong>and</strong> (a)<br />
showed higher adhesion, with values of 99.3 <strong>and</strong> 83.2 nN, respectively.<br />
As a general rule, roughness on two interacting surfaces reduces the area<br />
of contact between these surfaces, which would be expected to lead to<br />
a decrease in the measured adhesion. This could explain the increase<br />
in adhesion seen with surfaces (b) <strong>and</strong> (a) compared with the relatively<br />
rough surface (c). However, the least rough sample (a) has a lower adhesion<br />
than that of sample (b). If there were a simple relationship between<br />
the roughness of the surfaces <strong>and</strong> adhesion, this would not be expected<br />
to be the case. However, the relationship between roughness <strong>and</strong> adhesion<br />
is not easily defined <strong>and</strong> can be dependent upon the length scales<br />
of the roughness on the individual surfaces as well as the geometry of<br />
interaction of the surfaces. In some cases, if the length scales of the probe<br />
or asperities on the probe are similar to the roughness of the surface, it<br />
could be possible for an increased contact area to be achieved [13–16].<br />
4.8.1 Effect of Electropolishing of Steel Surfaces<br />
Electropolishing is a process of dissolving the very top layer of a metal<br />
surface to bring about an increase in the optical brightness or reflectivity<br />
of a surface by the reduction of surface roughness. One advantage of this<br />
over polishing of surfaces by mechanical means is the absence of defects<br />
in the surfaces due to scratches. As the characterisation of the roughness<br />
Mean adhesion (nN)<br />
120<br />
100<br />
80<br />
60<br />
40<br />
20<br />
0<br />
Polished steel Unpolished steel<br />
sample surface<br />
Corroded steel<br />
FIgURE 4.25 Summary of measured adhesion values between a calcium carbonate colloid<br />
probe <strong>and</strong> three steel sample surfaces.
4.8 CHARACTERIsATION OF METAL sURFACEs 131<br />
of surfaces at fine levels of detail is easily accomplished by AFM imaging,<br />
the monitoring of the polishing of surfaces is an obvious application.<br />
Abbot <strong>and</strong> co-workers used AFM both in air <strong>and</strong> in situ in liquid to<br />
monitor the changes in morphology over time of a stainless steel surface<br />
during electropolishing with an ionic liquid [17]. Roughness values,<br />
described by maximum z-height, were 597 <strong>and</strong> 88 nm for the unpolished<br />
<strong>and</strong> polished surfaces respectively. As the z-max figures for the unpolished<br />
surface were likely to be greater than would be expected owing to the surface<br />
not being flat, the authors also calculated roughness as a percentage<br />
difference in the flat scanning area versus the actual surface area of the<br />
sample. This yielded roughness values of 7.9% for the unpolished surface<br />
<strong>and</strong> 0.2% for the polished surface. It was also noted that a trough was seen<br />
to have been etched out at the boundary between the polished <strong>and</strong> unpolished<br />
areas, varying in depth from 1–2 �m, Figure 4.26.<br />
y / µm<br />
A<br />
70.0<br />
60.0<br />
50.0<br />
40.0<br />
30.0<br />
20.0<br />
10.0<br />
10.0 20.0 30.0 40.0 50.0 60.0<br />
x / µm<br />
Height/nm<br />
C<br />
500<br />
0<br />
–500<br />
–1000<br />
–1500<br />
–2000<br />
–2500<br />
Height/nm<br />
B<br />
500<br />
0<br />
–500<br />
–1000<br />
–1500<br />
–2000<br />
–2500<br />
10.0 20.0<br />
30.0<br />
40.0<br />
50.0 60.0<br />
70.0<br />
y / µm<br />
(x = 35.6µm)<br />
y / µm<br />
0.0<br />
10.0<br />
20.0<br />
30.0<br />
40.0<br />
50.0<br />
60.0<br />
70.0<br />
10.0<br />
20.0<br />
30.0<br />
40.0<br />
50.0<br />
60.0<br />
FIgURE 4.26 AFM images obtained at the transition between polished <strong>and</strong> non-<br />
polished areas of a stainless steel surface. (a) <strong>and</strong> (b) show the same areas in 2D <strong>and</strong> 3D<br />
modes. Unpolished surface is flat but has high roughness, whereas the polished surface<br />
is relatively smooth, but still shows large height variations. At the interface is an etched<br />
trench. (c) shows a line profile from the image, with the unpolished section on the left <strong>and</strong><br />
polished on the right. Reproduced from [17].<br />
x / µm
132 4. INvEsTIgATINg MEMbRANEs ANd MEMbRANE PROCEssEs<br />
Vignal et al. [18] also studied the electropolishing of steel surfaces<br />
using the AFM. They observed that under a specific set of conditions<br />
(in perchloric acid <strong>and</strong> monobutyl ether), a very regular network of hexagonal<br />
cells of approximately 100 nm periodicity <strong>and</strong> 10 nm depth was<br />
observed, Figure 4.27.<br />
This hexagonal pattern was described as being extremely sensitive<br />
to the applied voltage <strong>and</strong> was explained as being ‘prints of convective<br />
cells…localised in a resistive sub-layer of 100 nm thickness in the viscous<br />
shell’ [18].<br />
4.8.2 Corrosion of Metal Surfaces<br />
Almost all metal surfaces at some point in their lifetime either routinely<br />
or sporadically come into contact with corrosive media, whether<br />
it be aqueous solutions of high ionic strength or solutions of low pH. As<br />
a result, their behaviour when coming into contact with such chemicals<br />
<strong>and</strong> their ability to withst<strong>and</strong> corrosion play a significant part in their<br />
lifetime, durability, efficiency <strong>and</strong> appearance.<br />
One of the most common corrosive liquids present in the environment<br />
are aqueous NaCl solutions. Equipment used in the desalination industry,<br />
food processing <strong>and</strong> any equipment or structures in the maritime or nearmaritime<br />
environment will routinely come into contact with salt water,<br />
not to mention the corrosive effect on cars <strong>and</strong> roadside installations due<br />
500 nm<br />
FIgURE 4.27 Image obtained by AFM of regular hexagonal cells on steel surface<br />
formed after electropolishing, observed by Vignal et al. [18].
4.8 CHARACTERIsATION OF METAL sURFACEs 133<br />
to salt spray from roads during winter. Sanchez <strong>and</strong> colleagues [19] examined<br />
the early stages of corrosion of high-strength steel in dilute (50 mM)<br />
solutions of NaCl. The steels examined were of the type used for the reinforcing<br />
of concrete structures, which are prone to corrosion when exposed<br />
to salt water due to the porous nature of the concrete <strong>and</strong> consisting of<br />
ferrite <strong>and</strong> cementite. The authors scanned the same area of the steel surface<br />
for over 2 h <strong>and</strong> observed changes in the surface morphology due to<br />
interaction with the corrosive media. Snapshots of the sample at different<br />
times are shown in Figure 4.28.<br />
Initially (t � 0) the polishing marks <strong>and</strong> scratches are still visible, but<br />
disappear after 50 minutes, when the surface was observed to be increasingly<br />
rough as the growth of oxides on the surface occurred. After approximately<br />
50 min, <strong>and</strong> later, a series of ridges were observed to appear. The<br />
authors concluded that this was the lamellar structure of pearlite (the<br />
mixture of ferrite <strong>and</strong> cementite phases) appearing as the ferrite matrix<br />
was selectively attacked by the salt solution. The change in the surface<br />
roughness was monitored as a function of time, allowing the quantitative<br />
assessment of corrosion of the surface, Figure 4.29.<br />
Z: 24.1nm<br />
X: 5.0µm<br />
A B<br />
Z: 40.1nm<br />
X: 5.0µm<br />
C D<br />
Z: 21.6nm<br />
Z: 151.8nm<br />
X: 5.0µm<br />
X: 5.0µm<br />
FIgURE 4.28 Steel surface after varying lengths of exposure to 50 mM NaCl solution.<br />
(a) 0 min; (b) 30 min; (c) 50 min; (d) 2 h 15 min. Reproduced from [19].
134 4. INvEsTIgATINg MEMbRANEs ANd MEMbRANE PROCEssEs<br />
Surface roughnes (nm)<br />
1.6<br />
1.4<br />
1.2<br />
1<br />
0.8<br />
0.6<br />
0.4<br />
0.2<br />
0<br />
First<br />
stage<br />
Second<br />
stage<br />
y = –0.4907x 2 + 2.2621x–1.1032<br />
R 2 = 0.9906<br />
0 0.5 1 1.5 2<br />
Time (h)<br />
FIgURE 4.29 Change in roughness of steel surface over time during exposure to NaCl<br />
solution. Reproduced from [19].<br />
It is apparent that the roughness of the surface increased in two stages.<br />
In the first stage, no significant roughness increase was observed. After<br />
about 35 min, the roughness increased at a rapid rate, which could be fitted<br />
by a simple polynomial expression. The authors attributed the first stage to<br />
the formation of a thin layer of corrosion products, whilst the second stage<br />
was due to a selective attack of the ferrite phase of the steel.<br />
Other studies have concentrated on the corrosion of surfaces due to<br />
more aggressive corrosive agents. Wang et al. [20] studied the corrosion<br />
of stainless steel over a number of days by sulphuric acid microdrops <strong>and</strong><br />
thin films using a combination of AFM <strong>and</strong> XPS <strong>and</strong> observed the formation<br />
of pits <strong>and</strong> corrosion products. Solmaz <strong>and</strong> co-workers [21] examined<br />
the corrosion of mild steel by HCl solutions with <strong>and</strong> without the pres-<br />
ence of a corrosion inhibitory agent, using AFM to study changes in<br />
surface morphology in combination with a number of other techniques.<br />
When the surface was exposed to the HCl solution alone, corrosion pits<br />
were observed to form, which became both wider <strong>and</strong> deeper as the exposure<br />
time increased. When the surface was exposed to the HCl in combination<br />
with 10 mM 2-mercapto thiazoline, the surface was more uniform than<br />
surfaces exposed to HCl alone after 24 h exposure. For longer exposure<br />
times (approximately 120 h), a smoother surface was still evident when<br />
compared to treatment in the absence of 2-MT. A similar study by Qu et al.<br />
[22] examined the efficacy of a different additive (EDTA) on reduction of<br />
corrosion of steel surfaces by HCl.<br />
An interesting study was carried out by Valtiner et al. [23] into the effect<br />
of pH on acid dissolution of crystalline ZnO surfaces, which is often used<br />
as a protective coating for steels. It is the stability of the thin oxide covering<br />
that determines the corrosion resistance of zinc, <strong>and</strong> hence metal coated by<br />
it. The single crystalline surfaces were observed to consist of large flat sections<br />
several micrometres across, with step heights at the edges of 4–10 nm.<br />
When immersed in electrolytes of pH 11–5.5, no dissolution was observed.<br />
2.5
At pH 5.5, dissolution began to occur, but only at the step edges, not in the<br />
centre of the crystal planes. At pH values below this, as the pH decreased<br />
the dissolution rate increased, with a significant change in behaviour<br />
occurring below pH 4.0. Below this value, dissolution also began to occur<br />
at the flat crystal surfaces, although it was still most pronounced at the<br />
edges. It was suggested that this behaviour was due to the presence of<br />
oxygen in the step edges, which would be much more prone to attack by<br />
solution protons than the flat surfaces, which are positively charged at the<br />
higher pH values.<br />
A number of other studies have used AFM alongside other techniques<br />
to characterise a number of different coatings to protect metal surfaces<br />
from corrosion <strong>and</strong> wear, including TiN <strong>and</strong> ZrN thin films [24]; Sn–Ni<br />
<strong>and</strong> Sn–Cu alloy coatings [25, 26]; CrN films for enhanced wear resistance<br />
[27] <strong>and</strong> zirconia-based primers as coatings on aluminium [28].<br />
4.8.3 biofilms at Metal Surfaces<br />
4.9 CONCLUsIONs 135<br />
Yu <strong>and</strong> colleagues [29] examined the difference between nano- <strong>and</strong><br />
microcrystallization of stainless steel surfaces <strong>and</strong> the effect on the adhesion<br />
of biofilms. Using an AFM tip coated in a synthetic peptide to simulate<br />
a biofilm, adhesion measurements were made against nanocrystalline<br />
<strong>and</strong> microcrystalline surfaces. It was found that the measured adhesion<br />
in air was significantly reduced for the nanocrystalline surface compared<br />
with the microscrystalline surface. These observations were matched by<br />
similar results from the observations of attachment of whole bacterial<br />
cells to the same surfaces in bulk solution.<br />
4.9 COnCLUSIOnS<br />
The many benefits of AFM in investigations of membranes <strong>and</strong> membrane<br />
processes may be summarised as:<br />
1. Atomic force microscopy can determine the key properties of synthetic<br />
membranes: pore size distribution, surface morphology <strong>and</strong><br />
surface roughness, surface electrical properties <strong>and</strong> surface adhesion.<br />
2. Correspondence between surface pore dimensions from AFM <strong>and</strong><br />
MWCO is good. In addition, AFM gives surface pore size distribution.<br />
3. Operations in liquid <strong>and</strong> colloid probe techniques are particular<br />
advantages of AFM.<br />
4. AFM can establish the effects of changes in interactions over the surface<br />
of membranes, e.g. due to local morphology.<br />
5. AFM allows the visualisation of solute/membrane interactions.<br />
6. AFM is a very useful asset in assessing the properties of membranes<br />
during their development.
136 4. INvEsTIgATINg MEMbRANEs ANd MEMbRANE PROCEssEs<br />
Indeed, it is apparent that the capabilities of AFM are very closely matched<br />
to the knowledge requirements of membrane scientists <strong>and</strong> engineers.<br />
In addition, AFM is a very useful technique for investigating the<br />
interaction of process streams with the materials of construction of process<br />
equipment, including studies of particle adhesion <strong>and</strong> the effects of<br />
corrosion.<br />
ACknOWLEDgEMEnTS<br />
We thank all who have contributed to this research, especially Teodora<br />
Doneva, Daniel Johnson, Bob Lovitt, Martin Peer, Peter Williams, Chris<br />
Wright <strong>and</strong> Huabing Yin.<br />
LIST OF AbbREVIATIOnS<br />
BSA bovine serum albumin<br />
EDTA ethylenediaminetetraacetic acid<br />
MF microfiltration<br />
MT mercapto thiazoline<br />
MWCO molecular weight cut-off<br />
NF nanofiltration<br />
PSU polysulphone<br />
RO reverse osmosis<br />
SPEEK sulphonated poly (ether ether) ketone<br />
UF ultrafiltration<br />
XPS X-ray photoelectron spectroscopy<br />
Names of membranes are manufacturers’ codes.<br />
References<br />
[1] W.R. <strong>Bowen</strong>, N. <strong>Hilal</strong>, R.W. Lovitt, P.M. Williams, Atomic force microscope studies of<br />
membranes: surface pore structures of Cyclopore <strong>and</strong> Anopore membranes, J. Memb.<br />
Sci. 110 (1996) 233–238.<br />
[2] W.R. <strong>Bowen</strong>, N. <strong>Hilal</strong>, R.W. Lovitt, P.M. Williams, Visualisation of an ultrafiltration<br />
membrane by non-contact atomic force microscopy at single pore resolution, J. Memb.<br />
Sci. 110 (1996) 229–232.<br />
[3] W.R. <strong>Bowen</strong>, N. <strong>Hilal</strong>, R.W. Lovitt, C.J. Wright, A new technique for membrane characterisation:<br />
direct measurement of the force of adhesion of a single particle using an<br />
atomic force microscope, J. Memb. Sci. 139 (1998) 269–274.
REFERENCEs 137<br />
[4] W.R. <strong>Bowen</strong>, N. <strong>Hilal</strong>, R.W. Lovitt, C.J. Wright, Characterisation of membrane surfaces:<br />
direct measurement of biological adhesion using an atomic force microscope, J.<br />
Memb. Sci. 154 (1999) 205–212.<br />
[5] W.R. <strong>Bowen</strong>, T.A. Doneva, Atomic force microscopy characterisation of ultrafiltration<br />
membranes: correspondence between surface pore dimensions <strong>and</strong> molecular weight<br />
cut-off, Surf. Interface Anal. 29 (2000) 44–547.<br />
[6] W.R. <strong>Bowen</strong>, N. <strong>Hilal</strong>, R.W. Lovitt, A.O. Sharif, P.M. Williams, Atomic force microscope<br />
studies of membranes: force measurement <strong>and</strong> imaging in electrolyte solutions,<br />
J. Memb. Sci. 126 (1997) 77–89.<br />
[7] W.R. <strong>Bowen</strong>, T.A. Doneva, J.A.G. Stoton, The use of atomic force microscopy to quantify<br />
membrane surface electrical properties, Colloids Surf. A Physicochem. Eng. Asp.<br />
201 (2002) 73–83.<br />
[8] W.R. <strong>Bowen</strong>, T.A. Doneva, Atomic force microscopy studies of nanofiltration membranes:<br />
surface morphology, pore size distribution <strong>and</strong> adhesion, Desalination 129<br />
(2000) 63–172.<br />
[9] W.R. <strong>Bowen</strong>, N. <strong>Hilal</strong>, M. Jain, R.W. Lovitt, A.O. Sharif, C.J. Wright, The effects of electrostatic<br />
interactions on the rejection of colloids by membrane pores – visualisation<br />
<strong>and</strong> quantification, Chem. Eng. Sci. 54 (1999) 369–375.<br />
[10] W.R. <strong>Bowen</strong>, T.A. Doneva, H.B. Yin, Polysulphone-sulphonated poly (ether ether)<br />
ketone blend membranes: systematic synthesis <strong>and</strong> characterization, J. Memb. Sci. 181<br />
(2001) 253–263.<br />
[11] W.R. <strong>Bowen</strong>, T.A. Doneva, H.B. Yin, Separation of humic acid from a model surface<br />
water with PSU/SPEEK blend UF/NF membranes, J. Memb. Sci. 206 (2002) 417–429.<br />
[12] K. Al-Anezi, D.J. Johnson, N. <strong>Hilal</strong>, An atomic force microscope study of calcium<br />
carbonate adhesion to desalination process equipment: effect of anti-scale agent,<br />
Desalination 220 (2008) 359–370.<br />
[13] W.R. <strong>Bowen</strong>, R.W. Lovitt, C.J. Wright, Atomic force microscope studies of stainless steel:<br />
surface morphology <strong>and</strong> colloidal particle adhesion, J. Mater. Sci. 36 (2001) 623–629.<br />
[14] H.-J. Butt, B. Capella, M. Kapple, Force measurements with the atomic force microscope:<br />
technique, interpretation <strong>and</strong> applications, Surf. Sci. Rep. 59 (2005) 1–152.<br />
[15] T.S. Chow, Size-dependent adhesion of nanoparticles on rough substrates, J. Phys.<br />
Condens. Matt. 15 (2003) L83–L87.<br />
[16] Y.I. Rabinovich, J.J. Adler, A. Ata, R.K. Singh, B.M. Moudgil, Adhesion between nanoscale<br />
rough surfaces. II. Measurement <strong>and</strong> comparison with theory, J. Colloid Interface<br />
Sci. 232 (2000) 17–24.<br />
[17] A. Abbott, G. Capper, K.J. McKenzie, A. Glidle, K.S. Ryder, Electropolishing of stainless<br />
steels in a choline chloride based ionic liquid: an electrochemical study with surface<br />
characterisation using SEM <strong>and</strong> atomic force microscopy, Phys. Chem. Chem.<br />
Phys. (PCCP) 8 (2006) 4214–4221.<br />
[18] V. Vignal, J.C. Roux, S. Fl<strong>and</strong>rois, A. Fevrier, Nanoscopic studies of stainless steel electropolishing,<br />
Corr. Sci. 42 (2000) 1041–1053.<br />
[19] J. Sanchez, J. Fullea, C. Andrade, J.J. Gaitero, A. Porro, AFM study of the early corrosion<br />
of a high strength steel in a diluted sodium chloride solution, Corr. Sci. 50 (2008)<br />
1820–1824.<br />
[20] R. Wang, An AFM <strong>and</strong> XPS study of corrosion caused by micro-liquid of dilute sulfuric<br />
acid on stainless steel, Appl. Surf. Sci. 227 (2004) 399–409.<br />
[21] R. Solmaz, G. Kardas, M. Culha, B. Yazici, M. Erbil, Investigation of adsorption <strong>and</strong><br />
inhibitive effect of 2-mercaptothiazoline on corrosion of mild steel in hydrochloric<br />
acid media, Electrochim. Acta 53 (2008) 5941–5992.<br />
[22] Q. Qu, S. Jiang, W. Bai, L. Li, Effect of ethylene tetraacetic acid disodium on the corrosion<br />
of cold rolled steel in the presence of benzotriazole in hydrochloric acid,<br />
Electrochim. Acta 52 (2007) 6811–6820.
138 4. INvEsTIgATINg MEMbRANEs ANd MEMbRANE PROCEssEs<br />
[23] M. Valtiner, S. Borodin, G. Grundmeier, Stabilization <strong>and</strong> acidic dissolution mechanism<br />
of single-crystalline ZnO(0001) surfaces in electrolytes studied by in-situ AFM<br />
imaging <strong>and</strong> ex-situ LEED, Langmuir 24 (2008) 5350–5358.<br />
[24] D.F. Arias, Y.C. Arango, A. Devia, Study of TiN <strong>and</strong> ZrN thin films grown by cathodic<br />
arc technique, Appl. Surf. Sci. 253 (2006) 1683–1690.<br />
[25] B. Subramaniam, S. Mohan, S. Jayakrishnan, Structural, microstructural <strong>and</strong> corrosion<br />
properties of brush plated copper–tin alloys, Surf. Coat. Technol. 201 (2006)<br />
1145–1151.<br />
[26] B. Subramaniam, S. Mohan, S. Jayakrishnan, Selective area deposition of tin–nickel<br />
alloy coating – an alternative for decorative chromium plating, J. Appl. Electrochem.<br />
37 (2007) 219–224.<br />
[27] Y. Fu, X. Zhu, B. Tang, X. Hu, J. He, K. Xu, A.W. Batchelor, Development <strong>and</strong> characterization<br />
of CrN films by ion beam enhanced deposition, Wear 217 (1998) 159–166.<br />
[28] G. Gusmano, G. Montesperelli, M. Rapone, G. Padeletti, A. Cusma, S. Kaciulis, A.<br />
Mezzi, R. Di Maggio, Zirconia primers for corrosion resistant coatings, Surf. Coat.<br />
Technol. 242 (2007) 5822–5888.<br />
[29] B. Yu, E.M. Davis, R.S. Hodges, R.T. Irvin, D.Y. Li, Surface nanocrystallization of stainless<br />
steel for reduced biofilm adherence, Nanotechnology 19 (2008) 335101.
C H A P T E R<br />
5<br />
AFM <strong>and</strong> Development<br />
of (Bio)Fouling-Resistant<br />
Membranes<br />
<strong>Nidal</strong> <strong>Hilal</strong>, W. <strong>Richard</strong> <strong>Bowen</strong>,<br />
Daniel Johnson <strong>and</strong> Huabing Yin<br />
O U T L I N E<br />
5.1 Introduction 139<br />
5.2 Measurement of Adhesion of Colloidal Particles <strong>and</strong> Cells<br />
to Membrane Surfaces 141<br />
5.3 Modification of Membranes 144<br />
5.3.1 Modification of Membranes with Quaternary Ammonium Salts 144<br />
5.3.2 Membrane Characterisation 145<br />
5.3.3 (Bio)Adhesion Forces between a BSA-Functionalised<br />
Colloid Probe <strong>and</strong> Membrane Surface 151<br />
5.4 Modification of Membranes with Sulphonated Poly<br />
(Ether-Ether Ketone) Polymers 163<br />
5.5 Conclusions 168<br />
Acknowledgements 168<br />
Abbreviations <strong>and</strong> Symbols 169<br />
References 169<br />
5.1 IntroduCtIon<br />
Pressure-driven membrane processes are widely used in water treatment<br />
applications. However, membrane fouling occurs due to the deposition<br />
of suspended particulates <strong>and</strong> dissolved materials on the membrane<br />
Atomic Force Microscopy in Process Engineering 139<br />
© 2009, Elsevier Ltd
140 5. AFM AND DEvELOPMENT OF (BIO)FOULINg-REsIsTANT MEMBRANEs<br />
surface, restricting the efficiency of treatment processes which are dependent<br />
upon membrane filtration. Fouling becomes evident in the characteristic<br />
flux decline that results from increased resistance to flow due to<br />
the deposited material. Fouling analysis is difficult because the hydrodynamic<br />
aspects of concentration polarisation, as well as the physical<br />
chemistry of solute–solute <strong>and</strong> solute–membrane interactions must be<br />
considered.<br />
The formation of a fouling layer on the membrane surface, referred to<br />
as surface fouling, is the dominant fouling mechanism. Surface fouling<br />
involves the initial deposition of foulants of organic or inorganic origin<br />
on the membrane surface <strong>and</strong> the subsequent growth of a fouling layer<br />
that adversely influences membrane performance.<br />
Modification of the membrane surface can alter the surface chemistry<br />
<strong>and</strong> subsequently the surface charge properties of the membranes. The<br />
effectiveness of membranes significantly depends on their molecular <strong>and</strong><br />
structural architecture, i.e. functional group distribution (2D or threedimensional<br />
[3D] distribution <strong>and</strong> uniformity), shielding of functional<br />
groups <strong>and</strong> their accessibility, <strong>and</strong> membrane morphology. Therefore,<br />
membranes containing different functional groups <strong>and</strong> characterised by<br />
different structural architectures are of special interest for the advance of<br />
their application in different areas. In their development, profound structure<br />
characterisation is of particular importance in providing an insight<br />
into the relationships between membrane performance <strong>and</strong> structure.<br />
Changes in surface morphology (roughness, pore size <strong>and</strong> pore size distribution<br />
[PSD]) <strong>and</strong> surface charge due to modification of membrane<br />
surfaces have been reported by several researchers [1–3].<br />
The degree of hydrophobicity of both membrane <strong>and</strong> solutes determine<br />
the degree of fouling of membranes. It has been reported that<br />
protein fouling is greater for hydrophobic membranes than hydrophilic<br />
membranes [4]. However, the charge <strong>and</strong> hydrophobicity effects are<br />
functions of environmental conditions such as the salt concentration<br />
<strong>and</strong> the solution pH. The effect of charge decreases with increasing salt<br />
concentration due to a resultant decrease in the extent of the electrical<br />
double layer.<br />
Several authors have described natural organic matter (NOM) as<br />
one of the major membrane fouling agents in microfiltration, ultrafiltration<br />
<strong>and</strong> nanofiltration of surface water [5]. The chemical characteristics<br />
of the water in which the humic substances are dissolved, such as pH<br />
<strong>and</strong> ionic strength, are primary determinants in membrane fouling due<br />
to their effects on the conformational structure <strong>and</strong> charge of the humic<br />
substances.<br />
As bacterial surfaces contain proteins, polysaccharides <strong>and</strong> other<br />
biopolymers that can affect their adhesion to another surface, several
5.2 MEAsUREMENT OF ADHEsION OF COLLOIDAL PARTICLEs 141<br />
investigators have used different proteins to underst<strong>and</strong> the role of biopolymers<br />
<strong>and</strong> their physiochemical properties in bacterial adhesion [6, 7].<br />
In this chapter we will briefly summarise some of the literature pertaining<br />
to the adhesion of colloids <strong>and</strong> biocolloids to membrane surfaces<br />
as studied using AFM techniques. We will then concentrate on two more<br />
detailed examples in which the AFM has been used to characterise the<br />
physical <strong>and</strong> adhesive properties of membranes which have been developed<br />
specifically to reduce fouling.<br />
5.2 MEASurEMEnt of AdhESIon of ColloIdAl<br />
PArtIClES And CEllS to MEMbrAnE SurfACES<br />
The AFM has been used to measure adhesive forces between particles<br />
<strong>and</strong> various process surfaces. One important example is adhesion<br />
between particles, including both inorganic colloidal particles <strong>and</strong> bacterial<br />
cells, <strong>and</strong> filtration membranes. This interaction is of great importance<br />
when considering the fouling <strong>and</strong> biofouling of such surfaces.<br />
Particles adhere to the process membranes <strong>and</strong> reduce flow through the<br />
membrane, greatly reducing the filtration efficiency <strong>and</strong> working lifetime<br />
of the membranes. The process testing of new membranes is potentially<br />
expensive <strong>and</strong> time consuming. The quantification of adhesion forces<br />
between colloids <strong>and</strong> membranes can provide an important contribution<br />
to developing the theoretical prediction <strong>and</strong> optimisation <strong>and</strong> control<br />
of many engineering separation processes. As a result the development of<br />
AFM methods to quantify the adhesion of different materials to membranes<br />
of different compositions can potentially be very useful for membrane<br />
manufacturers <strong>and</strong> engineers [8]. When particle sizes are greater than the<br />
pore size in the absence of repulsive double layer interactions, such particles<br />
may plug the pores very effectively, leading to a catastrophic loss in<br />
filtration flux.<br />
Of the many established membrane characterisation techniques, only<br />
the colloid probe method can measure the adhesive forces between particles<br />
<strong>and</strong> membrane surfaces <strong>and</strong> hence allow prediction of the membrane<br />
fouling properties of the particles. In addition, the ability to make<br />
measurements in liquid allows the matching of observation conditions to<br />
those which occur in practice. The first demonstrations of the potential<br />
of the colloid probe technique to differentiate different membranes based<br />
upon the different adhesion properties was carried out by Brown <strong>and</strong> colleagues<br />
[8–12]. Two commercially available membranes with polymer<br />
molecular weights of 4 kDa were investigated. The first, ES 404, was made<br />
from a single polymer, polyethersulphone (PES). The second, XP 117, was<br />
made from a blend of different polymers created to have a potentially low
142 5. AFM AND DEvELOPMENT OF (BIO)FOULINg-REsIsTANT MEMBRANEs<br />
rate of membrane fouling. Measurements were made between polystyrene<br />
microspheres <strong>and</strong> these membranes in aqueous NaCl solutions at a 10 �2 M<br />
concentration <strong>and</strong> at a pH of 8.0. In Figure 5.1 plots of force normalised<br />
by the microsphere radius versus the displacement of the piezo in the<br />
direction normal to the membrane surface recorded whilst retracting the<br />
colloid probe away from the membrane surfaces are shown. At points A to<br />
A� the probe is in the region of constant compliance with the surface. For<br />
the region of the plot A� to B for the ES 404 membrane, a stretching of the<br />
probe <strong>and</strong>/or membrane occurs, due to the adhesive forces; the stretching<br />
continues at B to C, with adhesion force reducing as probe <strong>and</strong> membrane<br />
slowly separate, with final separation occurring at C. At C to D no<br />
net forces are acting between the probe <strong>and</strong> the surface, <strong>and</strong> this is taken as<br />
the point of zero force. The minimum force value (B) is taken as observed<br />
pull-off force F OFF <strong>and</strong> is a direct measure of the adhesive force. For the<br />
measurements presented in Figure 5.1, the values are 1.98 <strong>and</strong> 0.38 mN m �1<br />
for the ES 404 <strong>and</strong> XP 117 membranes, respectively – an approximately five-<br />
fold reduction. This demonstrates quantitatively that the membrane manufacturer<br />
has produced a membrane to which the test colloid attaches only<br />
weakly compared with a more conventional membrane. In other words the<br />
membrane is potentially low fouling in actual process applications.<br />
It is also worth noting that the adhesive interaction between the probe<br />
<strong>and</strong> the ES 404 membrane took place over a distance of approximately<br />
400 nm, most likely due to the stretching of the probe <strong>and</strong>/or the membrane<br />
surface. When the adhesion of a ‘cell probe’ (Figure 5.2) created<br />
F/R (mNm –1 )<br />
5<br />
4<br />
3<br />
2<br />
1<br />
0<br />
–1<br />
ES 404 membrane<br />
XP 117 membrane<br />
D C<br />
–2<br />
B<br />
–3<br />
1000 1200 1400 1600 1800 2000<br />
Piezo displacement (nm)<br />
fIgurE 5.1 Normalised retraction force versus displacement for a polystyrene colloid<br />
probe <strong>and</strong> two filtration membrane at pH 8.0, 10 �2 M NaCl. Adhesion against the XP 117<br />
membrane formulated for low fouling is much reduced compared with the more conventional<br />
ES 404 membrane.<br />
Aʹ<br />
A
5.2 MEAsUREMENT OF ADHEsION OF COLLOIDAL PARTICLEs 143<br />
with a single yeast cell to a silica surface was undertaken, the adhesive<br />
interaction also took place over a large distance [12], again most<br />
likely representing the stretching of the probe. Conversely, in studying<br />
the adhesion between systems of hard inorganic surfaces, the adhesive<br />
interactions take place over very short distances of the order of no more<br />
than a few nanometres [13–15]. This is a consideration for manufacturers<br />
of membranes when studying a heterogeneous range of materials. This<br />
behaviour is also of note for general colloid probe interactions. The deformation<br />
of soft surfaces in close proximity due to long-range <strong>and</strong> mechanical<br />
forces when in contact makes the determination of the actual surface<br />
separation <strong>and</strong> assignment of the point of zero distance problematic. For<br />
this reason many researchers plot only the displacement of the piezo,<br />
rather than the actual surface separation. This topic is discussed in more<br />
detail elsewhere in this book.<br />
Protein-coated colloid probes were used to compare the adhesive<br />
forces between silica colloids <strong>and</strong> bovine serum albumin (BSA)-coated<br />
silica spheres with the same membranes as described earlier [10]. BSAcoated<br />
silica probes demonstrated significant adhesion with both types<br />
of membranes, compared to silica probes, which had no measurable<br />
adhesion. The development of the colloid probe technique for the AFM<br />
as a sensor for quantifying the adhesive forces at membrane surfaces<br />
provides a relatively fast procedure for assessing the potential fouling of<br />
membrane surfaces by particles of different materials. In addition only<br />
fIgurE 5.2 SEM image of a yeast cell attached to the apex of an AFM microcantilever<br />
to create a ‘cell probe’.
144 5. AFM AND DEvELOPMENT OF (BIO)FOULINg-REsIsTANT MEMBRANEs<br />
small pieces of membrane are necessary for experiments to be undertaken.<br />
Ultimately the direct measurements of the AFM will help to assist<br />
in the development of new membranes with more fouling-resistant properties.<br />
(Further information on ES 404 <strong>and</strong> XP 117 is given in Chapter 4.)<br />
5.3 ModIfICAtIon of MEMbrAnES<br />
The development of high-performance membranes involves the<br />
selection of a suitable material <strong>and</strong> the formation of this material into a<br />
desired membrane structure. However, it is often necessary to modify the<br />
membrane material or the structure to enhance the overall performance<br />
of the membrane. The field of membrane technology is extremely broad,<br />
<strong>and</strong> the applicability of surface modification for different types of membranes<br />
is equally diverse. Polymeric membranes of nearly every variety<br />
have been utilised in the literature as substrate for polymer modification<br />
by the addition of another polymer.<br />
There are many techniques for the attachment of polymers to membrane<br />
surfaces. Some of these techniques, such as electrochemical <strong>and</strong><br />
�-irradiation, require continuous application of current or radiation<br />
throughout the polymerisation process. Others, such as UV-initiation,<br />
electron beam deposition, plasma treatment <strong>and</strong> silanisation, create reactive<br />
sites for subsequent formation of the polymer in situ, in general by<br />
free radical polymerisation. Whole macromolecules can be incorporated<br />
either by their cross-linking within the pores or by specific attachment of<br />
polymer chains to the membrane pore surfaces. Such modification methods<br />
are commonly applied to improve various membrane properties,<br />
including improvement of their surface-fouling resistance. Rendering the<br />
membrane surface more hydrophilic is a widely used method to reduce<br />
their tendency to become fouled.<br />
Modification by photo-initiated graft polymerisation with various<br />
hydrophilic monomers has been extensively applied to vary the hydrophilicity<br />
of the membrane surface. Acrylic acid, hydroxyethyl methacrylate<br />
(HEMA), poly(ethylene glycol) derivatives <strong>and</strong> vinyl pyrolidinone are<br />
examples of such hydrophilic monomers. Improvement in membrane fouling<br />
tendency has been reached after modification with such hydrophilic<br />
monomers [16–18].<br />
5.3.1 Modification of Membranes with Quaternary<br />
Ammonium Salts<br />
Quaternary 2-dimethyl-aminoethylmethacrylate (qDMAEMA) was<br />
chosen for the modification of PES <strong>and</strong> polyvinylidene fluoride (PVDF)<br />
membranes using photo-initiated graft copolymerisation method [1, 19, 20].
The antibacterial activity of initial unmodified (PES <strong>and</strong> PVDF) membranes<br />
as well as membranes modified with qDMAEMA polycations<br />
were evaluated against E. coli bacteria. After incubation with viable E. coli<br />
bacterial cells on each membrane, the growths of bacterial plaques were<br />
compared. Figure 5.3 shows PES membranes of which one is modified by<br />
incubation with 367 �g cm �2 qDMAEMA. The results showed that membrane<br />
samples with grafted qDMAEMA as well as ones modified with PEI<br />
possess a strong antimicrobial action towards E.coli bacteria.<br />
This antimicrobial activity was found to be independent with relation<br />
to the degree of modification (DM) of the membrane. The modified membranes<br />
possessing biocide properties could be potentially more resistant<br />
to (bio)fouling in water treatment applications.<br />
5.3.2 Membrane Characterisation<br />
5.3 MODIFICATION OF MEMBRANEs 145<br />
Membrane Surface Morphology<br />
Membrane surface topography <strong>and</strong> conformational changes due<br />
to modification were examined using AFM in contact mode. All AFM<br />
images were made in air at room temperature. Figure 5.4 presents 3D<br />
AFM images of initial PES membranes <strong>and</strong> modified membranes with<br />
different degrees of polymer grafting [8, 20].<br />
It can be seen that surface topography is markedly different in texture<br />
for unmodified membranes compared with the modified membranes,<br />
(a) (b)<br />
fIgurE 5.3 Viability of E.coli cells on the surface of initial PES (a) <strong>and</strong> modified<br />
with qDMAEMA membranes, DM � 367 �g cm �2 (b). Five millilitres of E. coli suspension<br />
(15 E. coli per ml) were filtered through both membranes.
146 5. AFM AND DEvELOPMENT OF (BIO)FOULINg-REsIsTANT MEMBRANEs<br />
25 µm<br />
12.5 µm 12.5 µm<br />
0 µm Z scale = 720 nm<br />
(a) (c)<br />
12.5 µm<br />
25 µm<br />
12.5 µm<br />
although both are characterised by the same nominal pore size. The surface<br />
of grafted membrane having the lowest DM (shortest polymer chain length)<br />
appeared to have a laterally inhomogeneous structure consisting of clusters.<br />
As the DM increased (leading to an increased graft chain length), the clusters<br />
became higher <strong>and</strong> the grafted chain gathered into larger clusters.<br />
Membrane Surface Analysis<br />
From AFM images, surface analysis was carried out using instrument<br />
software to obtain surface roughness parameters. The formation of<br />
the graft polymer layer on the membrane surface resulted in significant<br />
changes in surface morphology. Structural modifications become more<br />
pronounced with increasing DM. Figure 5.4 <strong>and</strong> Tables 5.1 <strong>and</strong> 5.2 present<br />
quantitative characteristics of the surface morphology of the initial<br />
<strong>and</strong> modified membranes derived from the AFM images [20].<br />
The same trend can be observed in the roughness values for both<br />
PES <strong>and</strong> PVDF membranes. A slight increase in the membrane roughness<br />
compared with the unmodified membrane was observed for membranes<br />
modified with the lowest degree of grafted polymer (202 �g cm �2 ).<br />
At such a DM, grafted chains could be accommodated within the pores.<br />
Further increases in DM led to a significant increase in surface roughness<br />
with a layer of grafted polymer forming its own porous structure.<br />
0 µm<br />
25 µm<br />
Z scale = 720 nm<br />
25 µm<br />
0 µm Z scale = 720 nm 0 µm Z scale = 720 nm<br />
(b) (d)<br />
fIgurE 5.4 High-resolution 3D AFM images of initial (a) <strong>and</strong> modified with<br />
qDMAEMA (b)–(d) PES membranes; (b) DM � 202 �g cm �2 , (c) DM � 367 �g cm �2 <strong>and</strong> (d)<br />
DM � 510 �g cm �2 .
5.3 MODIFICATION OF MEMBRANEs 147<br />
tAblE 5.1 AFM Measurements of surface Morphology of Initial <strong>and</strong><br />
Modified PEs Membranes with qDMAEMA.<br />
DM (�g cm �2 ) Ra (nm) rms (nm)<br />
Average<br />
Height (nm)<br />
Pore Size <strong>and</strong> Pore Size Distribution<br />
Pore size <strong>and</strong> PSD of each membrane were determined from AFMdetermined<br />
topography. The lognormal distribution was chosen to represent<br />
the pore size data for each of the membranes. This was found to give<br />
a good fit to the PSD. All of these distributions were fitted to lognormal<br />
distributions given by frequencies (%f ):<br />
1 dp<br />
% f � % fmax exp � ln<br />
2 X 2<br />
⎡<br />
2<br />
⎛ ⎞ ⎤<br />
⎢ ⎜ ⎥<br />
⎢ ⎜<br />
⎢ � ⎝⎜<br />
0 ⎠⎟<br />
⎥<br />
⎣<br />
⎦<br />
Maximum<br />
Range (nm)<br />
0 27.93 36.43 161.77 275.46<br />
202 32.26 42.91 250.09 378.4<br />
367 67.37 85.28 329.46 537.27<br />
tAblE 5.2 AFM Measurements of surface Morphology of Initial <strong>and</strong><br />
Modified PvDF Membranes with qDMAEMA.<br />
DM (�g cm �2 ) Ra (nm) rms (nm)<br />
Average<br />
Height (nm)<br />
Maximum<br />
Range (nm)<br />
0 91.52 114.38 402.07 699.97<br />
224 92.91 116.04 406.18 701.43<br />
346 101.68 127.42 433.97 834.53<br />
(5.1)<br />
where d p is the measured pore size, � the st<strong>and</strong>ard deviation of the measurements,<br />
%f max the maximum frequency <strong>and</strong> � 0 the modal value of d p.<br />
Mean pore sizes <strong>and</strong> PSDs of initial membranes determined from<br />
AFM images are shown in Table 5.3. This table shows that with a mean<br />
pore size of 0.535 � 0.082 �m, the PVDF membrane has slightly larger<br />
pores than the PES membrane (mean pore size 0.470 � 0.188 �m). PSDs<br />
are detailed in Figure 5.3, along with fits described by equation (5.1). The<br />
measured pore sizes are larger than the nominal pore size of 0.22 �m as<br />
specified by the manufacturers. The AFM data confirm that the pore
148 5. AFM AND DEvELOPMENT OF (BIO)FOULINg-REsIsTANT MEMBRANEs<br />
sizes of studied membranes are of approximately the same size. The<br />
topographical images give a clear perception of a notable difference in<br />
the surface morphology of the membranes used for the modification.<br />
A quantification of the surface parameters (Table 5.3) provides an insight<br />
into morphological particularities of these membranes which influence<br />
both the membrane separating properties <strong>and</strong> the process of modification<br />
by graft copolymerisation.<br />
The two membranes under study have notably different PSDs. It<br />
can be noted here that PVDF has a narrower PSD with pore sizes from<br />
0.336 to 0.68 �m compared with a PSD ranging from 0.219 to 0.948 �m in<br />
PES membranes. Moreover, these membranes significantly differ in surface<br />
roughness, with the PES membrane being smoother than the PVDF<br />
membrane. Regarding the AFM images, one might notice that the<br />
smoother surface allows for better contrast in pore observation, but more<br />
importantly the surface roughness is expected to have an influence on<br />
the graft copolymerisation.<br />
The rate of membrane modification was higher in the case of PES<br />
membrane than PVDF membrane [19]. It is impossible to associate the<br />
difference only with the contribution of surface morphology. It is well<br />
known that polysulphone <strong>and</strong> PES are intrinsically photo-active, undergoing<br />
bond cleavage with UV irradiation to produce free radicals even<br />
without the use of photo-initiators. PVDF is less photo-reactive than<br />
PES <strong>and</strong> produces less surface free radicals than PES. However, higher<br />
density of free radicals at the surface of more photo-reactive PES membranes<br />
also results in a higher probability of termination of chain growth<br />
<strong>and</strong> formation of cross-linked structures. These processes restricting an<br />
increase in the DM are competitive with respect to the chain growth.<br />
Since competitive processes, which enhance <strong>and</strong> decrease the amount of<br />
grafted polymer, occur simultaneously in the case of the photo-reactive<br />
polymer, the influence of surface morphology on graft copolymerisation<br />
should not be discarded. For the relatively rough surfaces, such as<br />
PVDF membrane, the decrease in UV-irradiation effectiveness <strong>and</strong> steric<br />
hindrance for polymer growth in narrow valleys are possible effects that<br />
may decrease the modification to some degree.<br />
tAblE 5.3 Parameters of Pore size <strong>and</strong> PsD Obtained from AFM Images for Initial<br />
PEs <strong>and</strong> PvDF Membranes.<br />
Pore size (�m) PSD parameters<br />
Membrane Mean Minimum Maximum X 0 (�m) %f max �<br />
PES 0.470 � 0.188 0.219 0.948 0.353 � 0.028 19.8 � 2.2 0.56 � 0.08
5.3 MODIFICATION OF MEMBRANEs 149<br />
Before detailed discussion of the quantitative characteristics of surface<br />
morphology, it is worth noting that the chosen lognormal pattern for<br />
PDS described by equation (5.1) gave a correlation coefficient of at least<br />
0.95 for all fitted curves. The most probable pore sizes estimated from<br />
fitted curves were very close to the mean pore diameter calculated from<br />
corresponding sets of pore sizes for the initial <strong>and</strong> modified membranes<br />
(Tables 5.4 <strong>and</strong> 5.5).<br />
According to Figure 5.5(a), the initial PES membrane has a very wide<br />
PSD with a � value of 0.56 �m. However, grafting of poly-qDMAEMA<br />
resulted in narrowing of the PSD <strong>and</strong> shifting the whole curve towards<br />
smaller pore sizes. As a result, mean pore size is gradually decreasing<br />
with the increase in the amount of poly-qDMAEMA grafted to the membrane<br />
surface. Significant improvement of the PSD was observed even<br />
for the modified PES membranes with the smallest DM (Table 5.4).<br />
Narrowing of the PSD occurred mostly due to the disappearance of<br />
large pores (larger than 0.6 �m). Taking into consideration that substantial<br />
narrowing of large pores dem<strong>and</strong>s higher quantities of grafted polymer<br />
tAblE 5.4 AFM Measurements of Pore size <strong>and</strong> PsD of Initial <strong>and</strong><br />
Modified PEs Membranes with qDMAEMA.<br />
Mean pore PSD parameters<br />
DM (�g/cm 2 ) size (�m) X 0 (�m) %f max �<br />
0 0.470 � 0.188 0.353 � 0.028 19.8 � 2.2 0.56 � 0.08<br />
202 0.337 � 0.098 0.278 � 0.010 41.4 � 4.4 0.32 � 0.04<br />
367 0.293 � 0.072 0.281 � 0.003 51.3 � 2.0 0.26 � 0.01<br />
510 0.100 � 0.083 0.075 � 0.004 26.9 � 2.9 0.40 � 0.05<br />
tAblE 5.5 AFM Measurements of Pore size <strong>and</strong> PsD of Initial<br />
<strong>and</strong> Modified PvDF Membranes with qDMAEMA.<br />
Mean pore PSD parameters<br />
DM (�g/cm 2 ) size (�m) X 0 (�m) %f max �<br />
0 0.535 � 0.082 0.555 � 0.003 51.2 � 1.7 0.14 � 0.01<br />
224 0.445 � 0.083 0.439 � 0.018 35.1 � 4.1 0.28 � 0.02<br />
346 0.334 � 0.079 0.297 � 0.013 44.3 � 5.7 0.30 � 0.01
150 5. AFM AND DEvELOPMENT OF (BIO)FOULINg-REsIsTANT MEMBRANEs<br />
Pore density %<br />
25<br />
20<br />
15<br />
10<br />
5<br />
0<br />
0<br />
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6<br />
(a) Pore size (µm)<br />
(c)<br />
Pore size (µm)<br />
Pore density %<br />
40<br />
30<br />
20<br />
10<br />
AFM data<br />
Fitted data<br />
AFM data<br />
Fitted data<br />
compared to smaller pores, it can be assumed that higher rates of polymer<br />
growth initiated at the walls of larger pores. As mentioned earlier,<br />
at the entrance of narrower pores, higher density of free radicals results<br />
in chain termination <strong>and</strong> consequently lower rate of polymer grafting.<br />
With time, when the PSD becomes more uniform, free radicals are eventually<br />
distributed across the membrane surface. This leads to a gradual<br />
decrease of all surface pores with PSD shifting to smaller sizes. With DM<br />
higher than 202 �m cm �2 , slight fluctuation in the width of PSD (�) was<br />
observed.<br />
It can be seen from Table 5.5 that similar behaviour is observed regarding<br />
changes in the surface morphology of the PVDF membrane as for the<br />
PES membrane. However, the unmodified PVDF membrane has a more<br />
uniform PSD than the PES membrane. With a mean pore size approximately<br />
0.54 �m, PSD of this membrane is characterised by a low value<br />
of �, �0.14 �m, compared with �0.56 �m for PES membranes. Although<br />
modification of PVDF membrane with grafted qDMAEMA also led to<br />
PSD shifting towards lower pore sizes, PSD was wider for the modified<br />
membranes compared with initial membrane.<br />
Pore density %<br />
Pore density %<br />
50<br />
40<br />
30<br />
20<br />
10<br />
30<br />
25<br />
20<br />
15<br />
10<br />
5<br />
AFM data<br />
Fitted data<br />
AFM data<br />
Fitted data<br />
0<br />
0<br />
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16<br />
(b)<br />
Pore size (µm)<br />
(d)<br />
Pore size (µm)<br />
fIgurE 5.5 PSDs of initial (a) <strong>and</strong> modified with qDMAEMA (b)–(d) PES membranes;<br />
(b) DM � 202 �g cm �2 , (c) DM � 367 �g cm �2 <strong>and</strong> (d) DM � 510 �g cm �2 .
Force (nN m –1 )<br />
Loading force<br />
(mN m –1 )<br />
52<br />
55<br />
61<br />
5.3 MODIFICATION OF MEMBRANEs 151<br />
Separation (nm)<br />
fIgurE 5.6 Representative approach (grey) <strong>and</strong> retraction (black) force curves of<br />
BSA-modified colloid probe to grafted membrane with qDMAEMA at different loading<br />
forces (pH 7).<br />
5.3.3 (bio)Adhesion forces between a bSA-functionalised<br />
Colloid Probe <strong>and</strong> Membrane Surface<br />
Adhesion force measurements were performed using the colloid probe<br />
technique. Colloid probes were prepared by immobilising modified<br />
microsphere with BSA to a tipless ultralever silicon cantilever.<br />
Examples of retraction curves measured at three different loading<br />
forces for a BSA-modified microsphere interacting with a PES membrane<br />
grafted with qDMAEMA, shown in Figure 5.6, demonstrate that the adhesion<br />
force (pull-off force) increases with the force applied via the colloid<br />
[20]. Similar findings have been observed in previous studies [7, 14]. The<br />
increase in the adhesion force is most likely to result from an increase in<br />
the number of chemical bonds formed between the biopolymers when the
152 5. AFM AND DEvELOPMENT OF (BIO)FOULINg-REsIsTANT MEMBRANEs<br />
surfaces are forced into closer contact using higher loading forces [7]. In<br />
addition the deformation of the membrane surface at high loading forces<br />
may increase the area of contact between the probe <strong>and</strong> the membrane<br />
surfaces. In order to exclude the influence of loading force in comparison<br />
of interactions force, the experimental results are presented by plotting the<br />
adhesion force versus the loading force. For the purpose of comparison,<br />
the adhesion force at loading force equal to 80 mN m �1 was measured,<br />
interpolated or extrapolated from the obtained experimental data.<br />
Interactions Forces between Initial Membranes <strong>and</strong> Functionalised<br />
Colloid Probes<br />
Effect of pH on Measured Adhesion Figure 5.7 shows the normalised<br />
adhesion force measurements versus the normalised loading force applied<br />
to the cantilever between a BSA-modified probe <strong>and</strong> initial PES membrane<br />
at different pH values. It can be seen that the adhesive force is highest at<br />
pH 5, followed by moderate values at pH 3 <strong>and</strong> the lowest at pH 7.<br />
First, at pH 7 both the BSA <strong>and</strong> the membrane surface are negatively<br />
charged [21], which provoke increases in the repulsive electrostatic interaction,<br />
leading to a reduction in adhesion. BSA is more hydrophobic at<br />
pH 5 than 7, which would be expected to increase adhesion. Secondly, the<br />
Adhesion force (mN m –1 )<br />
12<br />
10<br />
8<br />
6<br />
4<br />
2<br />
pH=3<br />
pH=5<br />
pH=7<br />
Loading force (mN m –1 0<br />
40 60 80 100 120 140<br />
)<br />
fIgurE 5.7 Relationship of adhesion <strong>and</strong> loading forces between initial PES <strong>and</strong> BSA<br />
probe in different pH solution. The dotted line is an extrapolated line.
5.3 MODIFICATION OF MEMBRANEs 153<br />
protein has a more exp<strong>and</strong>ed structure at pH values away from its isoelectric<br />
point (pI) of 4.8 [22]. Thus a steric repulsive interaction between<br />
the interacting surfaces will be higher at pH 3 than at pH 5, leading to<br />
reduced adhesion at pH 3. Finally, PES possesses a slight positive charge<br />
at pH 3 [21], provoking an electrostatic repulsive force.<br />
A linear relationship between adhesion <strong>and</strong> loading forces for interactions<br />
between BSA-modified probe <strong>and</strong> initial PVDF membrane at different<br />
pH is shown in Figure 5.8. A similar trend is observed to the initial<br />
PES membrane, with the adhesive force having the highest value at pH 5,<br />
whereas the lowest adhesion force was measured obtained at pH 7.<br />
The PVDF membrane possesses a negative charge in the pH range<br />
studied with its magnitude increasing with increased pH [23]. This<br />
explains the lower adhesion force between the BSA-modified probe<br />
<strong>and</strong> PVDF surface at pH 7 where both surfaces carry a negative charge.<br />
Increases in steric repulsion at pH 3 <strong>and</strong> in BSA hydrophobicity at pH 5<br />
resulted in higher adhesion at pH 5.<br />
Effect of Ionic Strength on Adhesion Forces The effect of ionic concentration<br />
on adhesion force is shown in Figure 5.9. Increased NaCl concentration<br />
led to an increase in the adhesion force between the BSA probe <strong>and</strong> the PES<br />
Adhesion force (mN m –1 )<br />
14<br />
12<br />
10<br />
8<br />
6<br />
4<br />
pH=3<br />
pH=5<br />
pH=7<br />
Loading force (mN m –1 2<br />
40 60 80 100 120 140 160 180 200<br />
)<br />
fIgurE 5.8 Relationship of adhesion <strong>and</strong> loading forces between initial PVDF <strong>and</strong> BSA<br />
probe in different pH solutions.
154 5. AFM AND DEvELOPMENT OF (BIO)FOULINg-REsIsTANT MEMBRANEs<br />
Adhesion force (mN m –1 )<br />
25<br />
20<br />
15<br />
10<br />
5<br />
1 mM NaCl<br />
10 mM NaCl<br />
100 mM NaCl<br />
Loading force (mN m –1 0<br />
60 80 100<br />
)<br />
120<br />
fIgurE 5.9 Relationship of adhesion <strong>and</strong> loading forces between initial PES <strong>and</strong> BSA<br />
probe in different ionic strength solution at pH � 5.8. The dotted line is an extrapolated line.<br />
membrane surface. The adhesion of the BSA probe to the initial unmodified<br />
PES increased by a factor of �2.4 across the concentration range 1–100 mM<br />
NaCl. This behaviour is qualitatively consistent with DLVO theory [24], as<br />
increasing the ionic strength compresses the thickness of the electrostatic<br />
double layer <strong>and</strong> should therefore reduce the repulsion force between the<br />
two surfaces, resulting in an increase in measured adhesion force.<br />
The adhesion force measured between a BSA-coated probe <strong>and</strong><br />
PVDF membrane surface at different ionic strength values is shown in<br />
Figure 5.10. As with the PES membrane, an increase in ionic strength<br />
led to an increase in the observed adhesion force. The adhesion force<br />
increased by a factor of 1.7 across the dissolved NaCl range of 1–100 mM.<br />
This is due to charge screening effects at higher ionic concentrations<br />
reducing the electrostatic repulsion, hence increasing the adhesion force<br />
between the interacting surfaces.<br />
Comparison of (Bio)Adhesion Force Measurements between PES<br />
<strong>and</strong> PVDF Initial Membranes<br />
Figure 5.11 shows the adhesion force for unmodified PES <strong>and</strong> PVDF<br />
membranes measured using BSA-coated colloid probes in solution at different<br />
pH values. A similar trend was observed for the adhesion force as<br />
a function of pH for both membranes. However, the PVDF membrane
Adhesion force (mN m –1 )<br />
10<br />
8<br />
6<br />
4<br />
2<br />
5.3 MODIFICATION OF MEMBRANEs 155<br />
0<br />
40 60 80 100 120 140 160<br />
Loading force (mN m –1 )<br />
1 mM NaCl<br />
10 mM NaCl<br />
100 mM NaCl<br />
fIgurE 5.10 Relationship of adhesion <strong>and</strong> loading forces between initial PVDF <strong>and</strong><br />
BSA probe in different ionic strength solutions at pH � 5.8.<br />
Adhesion force (mN m –1 )<br />
10<br />
8<br />
6<br />
4<br />
2<br />
0<br />
1 2 3 4 5 6 7 8 9<br />
pH<br />
PES<br />
PVDF<br />
fIgurE 5.11 Comparison of adhesion forces of BSA-modified probe to unmodified<br />
PES <strong>and</strong> PVDF membranes in various pH solutions.
156 5. AFM AND DEvELOPMENT OF (BIO)FOULINg-REsIsTANT MEMBRANEs<br />
showed a lower adhesion force than PES. This is most likely due to the<br />
PES membranes having a higher degree of hydrophobicity compared<br />
with PVDF. In addition, the PVDF membrane has a higher �-potential<br />
than the PES membrane [21, 23]. Thus a higher repulsive electrostatic<br />
force would be expected between the probe <strong>and</strong> the unmodified PVDF<br />
membrane than unmodified PES membrane.<br />
Effect of Ionic Strength A comparison of the adhesion force for both<br />
membranes as a function of ionic concentration is shown in Figure 5.12.<br />
In general a higher adhesion force is observed for PES membrane than<br />
PVDF membrane over the studied ionic strength range, except at ionic<br />
strength of 1 mM NaCl where adhesion force was small <strong>and</strong> no significant<br />
difference in adhesion of both membrane types is observed. At the<br />
highest ionic strength, the difference is the most dramatic. PES membrane<br />
is more hydrophobic than PVDF membrane. As a result a hydrophobic<br />
attractive force leads to higher adhesion force for PES membrane<br />
than PVDF membrane.<br />
Interactions Forces between Membranes Modified with qDMAEMA<br />
<strong>and</strong> BSA-Functionalised Colloid Probe<br />
Effect of pH The effect of solution pH on adhesion force for PES membranes<br />
modified with qDMAEMA is shown in Figure 5.13. An increase in<br />
adhesion force with increasing pH was observed for modified membranes<br />
with three different DMs <strong>and</strong> is approximately threefold at a loading force<br />
of 80 mN m �1 .<br />
The dominant contribution in the increase in adhesion force with increasing<br />
pH can be attributed to the electrostatic force between the positively<br />
charged modified membranes <strong>and</strong> the BSA probe. At pH 3 both surfaces<br />
(membrane <strong>and</strong> BSA probe) possess a positive charge resulting in greater<br />
repulsion at this pH <strong>and</strong> hence a lower adhesion than seen for other pH<br />
values. At pH 7 the modified PES membranes remains positively charged,<br />
whereas BSA carries a negative charge. This change in the nature of the<br />
electrostatic interaction leads to a greater observed adhesive force at pH 7.<br />
Figure 5.14 shows the results for the effect of pH on the measured adhesion<br />
force of BSA to PVDF membrane functionalised with qDMAEMA.<br />
An increase in adhesion force with increasing solution pH was again<br />
observed. The experimental results for PVDF grafted with different DMs<br />
reveal the same trend. Taking into consideration the amphoteric behaviour<br />
of BSA, which leads to a change in protein charge with changing<br />
pH, the BSA possesses a positive charge at pH values below the pI <strong>and</strong><br />
a negative charge at pH values above the isoelectric point. This explains<br />
the relatively high adhesion at pH 7 where both the BSA <strong>and</strong> the membrane<br />
surface are oppositely charged. Thus, an attractive electrostatic force<br />
between the interacting surfaces results in higher adhesion.
Adhesion force (mN m –1 )<br />
20<br />
15<br />
10<br />
5<br />
5.3 MODIFICATION OF MEMBRANEs 157<br />
0<br />
0.1 1 10 100<br />
Ionic strength (mM)<br />
PES<br />
PVDF<br />
fIgurE 5.12 Comparison of adhesion forces of BSA-modified probe to initial PES <strong>and</strong><br />
PVDF membranes in various ionic strength solutions.<br />
Adhesion force (mN m –1 )<br />
8<br />
7<br />
6<br />
5<br />
4<br />
3<br />
2<br />
1<br />
20 40 60 80 100 120 140<br />
Loading force (mN m –1 )<br />
pH=3<br />
pH=5<br />
pH=7<br />
fIgurE 5.13 Relationship of adhesion <strong>and</strong> loading forces between BSA probe <strong>and</strong> PES<br />
membrane modified with qDMAEMA (DM � 510 �g cm �2 ) at different pH solution. The<br />
dotted line is an extrapolated line.
158 5. AFM AND DEvELOPMENT OF (BIO)FOULINg-REsIsTANT MEMBRANEs<br />
Adhesion force (mN m –1 )<br />
3.5<br />
3.0<br />
2.5<br />
2.0<br />
1.5<br />
1.0<br />
0.5<br />
Loading force (mN m –1 0.0<br />
60 70 80<br />
)<br />
90<br />
pH=3<br />
pH=5<br />
pH=7<br />
fIgurE 5.14 Relationship of adhesion <strong>and</strong> loading forces between BSA probe <strong>and</strong><br />
PVDF membrane modified with qDMAEMA (DM � 530 �g cm �2 ) in different pH solution.<br />
Steric interaction could be another factor which plays a role in reducing<br />
the adhesion force observed with both membranes at pH 3. The overlap<br />
of the exp<strong>and</strong>ed protein layer <strong>and</strong> the modification of polymer chain<br />
of qDMAEMA, which is fully dissociated at this pH, enhance the repulsive<br />
interactions between the colloid probe <strong>and</strong> the surface of the modified membrane.<br />
In addition, at pH 5 there is practically no electrostatic interaction<br />
because this pH is close to the isoelectric point of the protein (the net charge<br />
is around zero) <strong>and</strong>, therefore, the electrostatic repulsion is negligible. As<br />
at pH 5 the protein has no net charge <strong>and</strong> so the structure of the protein is<br />
more compact <strong>and</strong> it has hydrophobic properties, such interaction may lead<br />
to high adhesion at pH 5 than at pH 3.<br />
Effect of Ionic Strength Normalised adhesion forces versus normalised<br />
loading forces for interactions involving the modified PES membrane are<br />
shown in Figure 5.15 at different solution ionic strengths. An increase in<br />
adhesion force with increased ionic strength is evident.<br />
The normalised adhesion force measured between BSA versus PVDFmodified<br />
membrane with qDMAEMA was examined as a function of<br />
ionic strength by varying the NaCl concentration (Figure 5.16). The<br />
adhesion force was also found to increase with increasing solution ionic<br />
strength. This is due to the compression of the double layer, which leads<br />
to the increase in adhesion force with increasing ionic strength.
Adhesion force (mN m –1 )<br />
10<br />
8<br />
6<br />
4<br />
2<br />
5.3 MODIFICATION OF MEMBRANEs 159<br />
1 mM NaCl<br />
10 mM NaCl<br />
100 mM NaCl<br />
Loading force (mN m –1 0<br />
0 20 40 60 80 100 120<br />
)<br />
fIgurE 5.15 Relationship of adhesion <strong>and</strong> loading forces between BSA probe <strong>and</strong><br />
modified membrane with qDMAEMA (DM � 510 �g cm �2 ) in different ionic strength solutions<br />
at pH � 5.8. The dashed line is an extrapolated line.<br />
Comparison of (Bio)Adhesion Force Measurements Between<br />
Unmodified Membranes <strong>and</strong> Membranes Modified with qDMAEMA<br />
Effect of pH The normalised adhesion forces at the same loading force<br />
(loading force � 80 mN m �1 ) for initial <strong>and</strong> modified PES membranes<br />
with qDMAEMA are shown in Figure 5.17. It is clear that the unmodified<br />
membrane has a higher adhesion force than the modified membranes at<br />
pH 3 <strong>and</strong> 5, whereas the lowest adhesion was observed at pH 7. In addition,<br />
at pH 3 modified membrane with the lowest DM exhibits the lowest<br />
adhesion. At pH 5 modified membrane with DM of 510 �g cm �2 showed<br />
the lowest adhesion force whereas both membranes show almost the<br />
same adhesion at pH 7.<br />
The BSA-functionalised probe <strong>and</strong> both modified <strong>and</strong> unmodified<br />
membranes all possess a positive surface charge at pH 3. However, the<br />
qDMAEMA-functionalised membrane possesses a markedly greater<br />
charge than the unmodified PES membrane. The resulting greater electrostatic<br />
repulsion accounts for the greater adhesive forces observed with<br />
the initial unmodified PES membrane than with the modified membrane.<br />
When the surrounding solution is at pH 7, both the initial PES membrane<br />
<strong>and</strong> the BSA possess negative surface charges, with the qDMAEMA<br />
membrane still carrying a positive charge. This accounts for the modified<br />
membrane having a greater adhesive force with the BSA-coated probe<br />
than the unmodified membrane.
160 5. AFM AND DEvELOPMENT OF (BIO)FOULINg-REsIsTANT MEMBRANEs<br />
Adhesion force (mN m –1 )<br />
7<br />
6<br />
5<br />
4<br />
3<br />
2<br />
1<br />
0<br />
40 60 80 100 120 140<br />
Loading force (mN m –1 )<br />
1 mM NaCl<br />
10 mM NaCl<br />
100 mM NaCl<br />
fIgurE 5.16 Relationship of adhesion <strong>and</strong> loading forces between BSA probe <strong>and</strong><br />
PVDF membrane modified with qDMAEMA (DM � 530 �g cm �2 ) in different ionic strength<br />
solutions at pH � 5.8.<br />
Adhesion force (mN m –1 )<br />
12<br />
10<br />
8<br />
6<br />
4<br />
2<br />
0<br />
1 2 3 4 5 6 7 8 9<br />
pH<br />
Initial PES<br />
Modified PES DM = 202 μg cm –2<br />
Modified PES DM = 367 μg cm –2<br />
Modified PES DM = 510 μg cm –2<br />
fIgurE 5.17 Comparison of adhesion forces of BSA-modified probe to initial <strong>and</strong><br />
modified with qDMAEMA membranes in various pH solutions.
5.3 MODIFICATION OF MEMBRANEs 161<br />
Figure 5.18 shows the adhesion forces between initial <strong>and</strong> modified<br />
PVDF membranes <strong>and</strong> the BSA-functionalised probe. At pH 3 <strong>and</strong> 5 the<br />
initial PVDF membrane showed a higher adhesion than the modified<br />
PVDF membrane. At the highest pH at which measurements were made,<br />
the adhesion for the unmodified membrane was lower than for the modified<br />
membranes, except for the membrane with the highest DM.<br />
Effect of Ionic Strength The adhesion force measured between BSAcoated<br />
colloid <strong>and</strong> modified PES membranes is shown in Figure 5.19. The<br />
adhesion force for all the investigated membranes was found to increase<br />
with increasing solution ionic strength. The modified membrane with<br />
DM 367 �g cm �2 demonstrated the lowest adhesion force followed by the<br />
initial membrane, which shows an increase in adhesion force compared<br />
to modified membrane with medium DM.<br />
Figure 5.20 shows the adhesion results for initial <strong>and</strong> grafted PVDF<br />
membranes using BSA probes at different solution ionic strengths. An<br />
increase in adhesion force with increasing ionic strength was observed<br />
for both unmodified <strong>and</strong> modified membranes. Again a similar result to<br />
PES membranes is noticed here where adhesion force has its lowest value<br />
for modified PVDF with DM 346 �g cm �2 except at ionic strength of 1 mM<br />
NaCl whereas grafted membrane with DM 530 �g cm �2 shows a slightly<br />
lower adhesion than modified with DM 346 �g cm �2 .<br />
Adhesion force (mN m –1 )<br />
12<br />
10<br />
8<br />
6<br />
4<br />
2<br />
Initial PVDF<br />
Modified PVDF DM = 224 μg cm –2<br />
Modified PVDF DM = 348 μg cm –2<br />
Modified PVDF DM = 530 μg cm –2<br />
0<br />
1 2 3 4 5<br />
pH<br />
6 7 8 9<br />
fIgurE 5.18 Comparison of adhesion forces of BSA-modified probe to initial <strong>and</strong><br />
modified with qDMAEMA membranes in various pH solutions.
162 5. AFM AND DEvELOPMENT OF (BIO)FOULINg-REsIsTANT MEMBRANEs<br />
Adhesion force (mN m –1 )<br />
60<br />
50<br />
40<br />
30<br />
20<br />
10<br />
Initial PES<br />
Modified PES DM = 202 μg cm –2<br />
Modified PES DM = 367 μg cm –2<br />
Modified PES DM = 510 μg cm –2<br />
0<br />
0.1 1 10<br />
Ionic strength (mM)<br />
100<br />
fIgurE 5.19 Comparison of adhesion forces of BSA-modified probe to initial <strong>and</strong><br />
modified with qDMAEMA membranes in solutions of increasing NaCl concentration.<br />
Adhesion force (mN m –1 )<br />
12<br />
10<br />
8<br />
6<br />
4<br />
2<br />
0<br />
0.1 1 10 100<br />
Ionic strength (mM)<br />
Initial PVDF<br />
Modified PVDF DM = 224 μg cm –2<br />
Modified PVDF DM = 348 μg cm –2<br />
Modified PVDF DM = 530 μg cm –2<br />
fIgurE 5.20 Comparison of adhesion forces of BSA-modified probe to initial <strong>and</strong><br />
modified with qDMAEMA membranes in various ionic strength solutions.
5.4 MODIFICATION OF MEMBRANEs wITH sULPHONATED 163<br />
The lower adhesion for modified PVDF membranes compared to the<br />
initial PVDF especially at DM 348 �g cm �2 suggests that the hydrophobic<br />
interactions are playing the dominant role where the higher hydrophobicity<br />
of initial membranes results in increasing its adhesion compared to<br />
modified membranes.<br />
5.4 ModIfICAtIon of MEMbrAnES wIth<br />
SulPhonAtEd Poly(EthEr-EthEr KEtonE) PolyMErS<br />
One of the most widely used materials in the manufacture of filtration<br />
membranes is polysulphone, due to its excellent mechanical, thermal <strong>and</strong><br />
chemical stability. Unfortunately a high degree of hydrophobicity possessed<br />
by unmodified polysulphone membranes renders them prone to<br />
fouling by a wide range of solutes. To improve membrane effectiveness by<br />
reducing fouling, <strong>and</strong> particularly biofouling, hydrophilic polymers may<br />
be incorporated into membranes or the membranes modified by the addition<br />
of extra functional groups by, for instance, sulphonation. The addition<br />
of charge bearing groups to the membrane surface will not only decrease<br />
the degree of hydrophobicity of the membrane but can also cause increased<br />
rejection of particles <strong>and</strong> solutes of identical charge [25].<br />
Poly(ether-ether ketone) (PEEK or poly(oxy-1,4-phenyleneoxy-1,4-<br />
phenylcarbonyl-1,4-phenylene) is a very chemically stable polymer,<br />
soluble only in strong acids. This includes concentrated sulphuric or<br />
chlorsulphonic acid, which yields a sulphonated PEEK (SPEEK) [26, 27].<br />
Studies of solubility of SPEEK suggest that it is more hydrophilic than<br />
merely sulphonated polysulphone [28].<br />
Studies have been carried out to investigate the effectiveness of using<br />
SPEEK as a charged polymer additive to polysulphone membranes to<br />
not only reduce fouling by the introduction of charged groups, but also<br />
as a pore-forming agent due to its hydrophilicity [29, 30]. In particular<br />
here we will report the use of AFM in the characterisation of the effects<br />
of the SPEEK additives on surface morphology <strong>and</strong> fouling resistance.<br />
All membranes were characterised using AFM-obtained topographies.<br />
In Figure 5.21 example images of a plain polysulphone membrane <strong>and</strong><br />
a membrane of polysulphone blended with 5% SPEEK are shown. Data<br />
obtained from AFM images, including average pore size (r p), root mean<br />
square (rms) roughness <strong>and</strong> surface porosity, are summarised in Table 5.6.<br />
Little variation is seen in rms roughness, although there is a trend for<br />
decreasing roughness with higher SPEEK content. Porosity variation<br />
follows a similar trend to that calculated from permeability of the membrane<br />
to water [29], but of a much smaller magnitude, suggesting that the<br />
increase in water permeability as SPEEK content increases is only partly
164 5. AFM AND DEvELOPMENT OF (BIO)FOULINg-REsIsTANT MEMBRANEs<br />
300<br />
(a)<br />
Å<br />
2.00<br />
1.00<br />
0.00<br />
200<br />
Å<br />
100<br />
0 0<br />
100<br />
Å<br />
200<br />
300<br />
Å<br />
4.00<br />
3.00<br />
2.00<br />
1.00<br />
0.00<br />
200<br />
Å<br />
100 100<br />
due to increase in the surface porosity. Changes in the thickness <strong>and</strong>/or<br />
the structure of the porous layer of the membrane may account for this.<br />
Force measurements were also carried out between �4-�m silica<br />
spheres <strong>and</strong> membranes. Figure 5.22 shows examples of typical data for<br />
force measurements taken from polysulphone membranes, with <strong>and</strong> without<br />
SPEEK present in 0.01 M NaCl. For the non-SPEEK membrane (P-P) the<br />
colloid probe snaps-in to the membrane surface, which indicates the presence<br />
of long-range attractive forces, sufficient in magnitude to overcome<br />
the restoring force on the cantilever. This type of attraction leads to fouling<br />
300<br />
0 0<br />
fIgurE 5.21 AFM images of polysulphone membranes obtained with contact mode in<br />
air. (a) Polysulphone membrane <strong>and</strong> (b) polysulphone/SPEEK membrane (5% SPEEK prepared<br />
at 20°C). Roughness values were obtained from 2 � 2 �m images.<br />
tAblE 5.6 summary of AFM Characterisation Data for sPEEK Modified<br />
Polysulfone Membranes.<br />
(b)<br />
P-P S0.5-20 S2-20 S5-20 S2-10 S2-60<br />
r p (nm) 0.93 � 0.15 0.89 � 0.22 0.93 � 0.15 0.73 � 0.18 1.05 � 0.22 0.86 � 0.26<br />
rms roughness<br />
(nm)<br />
3.2 4.2 3.5 2.5 3.1 2.1<br />
Porosity, e (%) 6.0 9.0 14.1 16.9 15.1 12.8<br />
Snap-in events<br />
(out of 9)<br />
F off, F/R<br />
(mN/m)<br />
9 4 3 0 1 4<br />
28.5 � 4.3 10.0 � 14.8 3.9 � 1.6 0.75 � 2.7 � 2.8 9.6 � 5.3<br />
Å<br />
200<br />
300
5.4 MODIFICATION OF MEMBRANEs wITH sULPHONATED 165<br />
FIR/mN m –1<br />
(a)<br />
FIR/mN m –1<br />
(b)<br />
2<br />
1<br />
0<br />
–1<br />
5<br />
0<br />
–5<br />
–10<br />
–15<br />
–20<br />
–25<br />
–30<br />
0 20 40 60<br />
Distance/nm<br />
–2<br />
0<br />
Approach<br />
Retraction<br />
20<br />
Distance/nm<br />
Approach<br />
Retraction<br />
fIgurE 5.22 Example normalised force versus distance curves against (a) P-P type<br />
membrane <strong>and</strong> (b) S5-20 membrane.<br />
of the membrane. When the probe is retracted away from the membrane<br />
surface, a large attractive force is again measured. The difference between<br />
the attractive force at its greatest magnitude <strong>and</strong> the force measured when<br />
the probe has no net forces acting upon it (its free level) is the measured<br />
adhesion force, F off. In contrast, for the S5-20 SPEEK membrane, no snapin<br />
is observed. This is most likely due to the silica probe <strong>and</strong> membrane<br />
surface carrying charges which are identical in sign. When the particle<br />
is retracted, only a moderate adhesive force was measured. Data for the<br />
40<br />
60
166 5. AFM AND DEvELOPMENT OF (BIO)FOULINg-REsIsTANT MEMBRANEs<br />
mean observed adhesion forces <strong>and</strong> the number of snap-in events measured<br />
for each membrane are also tabulated in Table 5.6, based on the measurements<br />
taken from nine different points on each membrane surface.<br />
All measurements for P-P membrane contained snap-in events, but this<br />
decreased as the SPEEK content of the membrane increased. For the S5-20<br />
membrane, containing 5% SPEEK, no snap-in event was measured at all.<br />
As the SPEEK content of the membranes is increased, a substantive <strong>and</strong><br />
systematic decrease in mean observed adhesion forces is seen, being a factor<br />
of almost 40 between P-P <strong>and</strong> S5-20 membranes. These trends together<br />
show the profound influence on the surface properties of the membranes<br />
of the incorporation of small amounts of SPEEK. It is also notable that the<br />
presence of snap-in events in some cases in the S-series membranes, except<br />
for S5-20, <strong>and</strong> the large st<strong>and</strong>ard deviation values show that some variability<br />
in the surface properties of the membranes exist.<br />
AFM has also been used to help in the assessment of fouling by<br />
humic acid (HA) of SPEEK-modified polysulphone membranes [30].<br />
HAs are heterogeneous materials, containing three main types of functional<br />
groups: carboxylic acids, phenolic acids <strong>and</strong> methoxycarbonyls.<br />
They are mostly negatively charged for pH values above pH 2.8 [31].<br />
Measurements were carried out with the S5-20 SPEEK membrane, <strong>and</strong><br />
an aromatic PES membrane (ES404), which were chosen as a comparable<br />
commercially available membrane [30].<br />
The effects of a deposited HA layer on membrane filtration will<br />
depend to some extent on its physical state. Images of membranes<br />
ES404 <strong>and</strong> S5-20 prior to use in filtering HA from water are shown in<br />
Figure 5.23(a). Before use, the S5-20 membrane has a rougher surface,<br />
visible in the images, <strong>and</strong> also indicated by a greater rms roughness.<br />
Figure 5.23(b) shows images of the two same membranes after 2 h filtering<br />
HA from model water. For the ES404 membrane the deposit is compact<br />
with occasional larger spheroidal aggregates. The rms surface roughness<br />
was greater by a factor of approximately 1.9 compared to before filtration.<br />
For the S5-20 SPEEK membrane, the deposit is much greater due to<br />
a higher flux [30]. The deposit is also less compact <strong>and</strong> more irregular than<br />
that seen for the ES404 membrane, <strong>and</strong> the rms surface roughness has<br />
increased by a factor of approximately 5.6 compared to before filtration.<br />
Both membranes were rinsed subsequent to filtration. Recovered membranes<br />
are shown in Figure 5.23(c). For the ES404 membrane, the surface<br />
roughness is similar to the fouled membrane, suggesting that rinsing has<br />
had little effect on removing the fouling aggregates. For the S5-20 membrane,<br />
however, the roughness value is greatly reduced from the fouled<br />
membrane, but has not quite returned to the value seen for the unused<br />
membrane. This suggests that most, but not all, of the fouling material<br />
has been removed by simple rinsing.
Å<br />
200<br />
100<br />
0<br />
5.4 MODIFICATION OF MEMBRANEs wITH sULPHONATED 167<br />
ES404: R p-v 26.6 nm; R rms 1.6 nm S5-20: R p-v 33.7 nm; R rms 2.5 nm<br />
2<br />
2<br />
2<br />
1.5<br />
1.5<br />
1.5<br />
1.5<br />
µm<br />
1<br />
0.5 0.5<br />
1<br />
µm<br />
1<br />
µm<br />
0.5<br />
0.5<br />
1<br />
µm<br />
(a) 0 0<br />
0 0<br />
Å<br />
300<br />
200<br />
100<br />
0<br />
2<br />
1.5<br />
Å<br />
200<br />
100<br />
0<br />
ES404: Rp-v 39.9 nm; Rrms 3.0 nm S5-20: R 124.0 nm; R 14.1 nm<br />
p-v rms<br />
Å<br />
1000<br />
1 1<br />
0.5 0.5<br />
(b) 0 0<br />
Å<br />
300<br />
2<br />
200<br />
100<br />
0<br />
(c)<br />
1.5<br />
µm<br />
µm<br />
1<br />
ES404: R p-v 32.6 nm; R rms 2.9 nm<br />
0.5 0.5<br />
0 0<br />
1<br />
1.5<br />
2<br />
800<br />
400<br />
2<br />
0<br />
Å<br />
400<br />
300<br />
200<br />
100<br />
0<br />
2<br />
1.5<br />
µm µm<br />
1.5<br />
2<br />
1.5<br />
µm µm<br />
1<br />
0.5 0.5<br />
0 0<br />
2<br />
1.5<br />
1<br />
µm<br />
S5-20: R p-v 42.0 nm; R rms 3.9 nm<br />
1.5<br />
1 1<br />
µm<br />
0.5 0.5<br />
0 0<br />
fIgurE 5.23 AFM images of ES404 <strong>and</strong> S5-20 membranes: (a) fresh membrane,<br />
(b) membrane after 2 h of filtering HA in water, <strong>and</strong> (c) recovered membrane after rinsing.<br />
R p–v � peak to valley roughness; R rms � rms roughness.<br />
2<br />
2
168 5. AFM AND DEvELOPMENT OF (BIO)FOULINg-REsIsTANT MEMBRANEs<br />
The polymers from which the ES404 <strong>and</strong> S5-20 membranes are primarily<br />
constructed, PES <strong>and</strong> polysulphone, respectively, are similar in physical<br />
<strong>and</strong> chemical characteristics. However, the blending of SPEEK into the S5-<br />
20 has resulted in a greater surface charge, greater porosity <strong>and</strong> a greater<br />
roughness. These properties have also altered the way in which the HA<br />
deposits are structured. The much less compact nature of the HA layer on<br />
the S5-20 membrane has resulted in easier rinsing <strong>and</strong> cleaning, despite a<br />
higher deposition of material at the membrane surface.<br />
Membranes containing blended SPEEK, such as the S5-20 membrane,<br />
show a combination of permeability, rejection properties <strong>and</strong> low adhesion<br />
characteristics. Attraction <strong>and</strong> adhesion to the membranes by silica<br />
colloids have been shown to be much reduced. In addition fouling by HA<br />
has been shown to be easily removed by simple rinsing of the membrane<br />
surface due to the open <strong>and</strong> un-compact structure of the deposited layers.<br />
Further details of the development of PSU/SPEEK membranes are<br />
given in Chapter 4.<br />
5.5 ConCluSIonS<br />
In this chapter we have discussed the contribution that AFM may make<br />
in the development of biofouling-resistant filtration membranes. The ability<br />
to image the 3D topography of the membrane surface allows quantitative<br />
assessment of both roughness <strong>and</strong> PSD. Measurement of roughness before<br />
<strong>and</strong> after the use of a membrane gives a quantitative measure of the degree<br />
of fouling on the membrane <strong>and</strong> allows comparison between different<br />
membranes. The PSD gives an indication of porosity, which complements<br />
observations which may be made using other techniques. In addition the<br />
ability of the AFM to use the colloid probe technique to directly assess the<br />
interaction forces between colloids <strong>and</strong> membrane surfaces under different<br />
conditions is of great use when it comes to assessing the capability of<br />
newly developed membranes to reject fouling particulates.<br />
In the future AFM has the capabilities to make a significant contribution<br />
in the development of new membrane surfaces <strong>and</strong> their properties, both<br />
by allowing characterisation of surface morphology <strong>and</strong> by measuring the<br />
physical interactions between membranes <strong>and</strong> potential fouling materials.<br />
ACKnowlEdgEMEntS<br />
We thank all who have contributed to this research, especially Laila<br />
Al-Khatib <strong>and</strong> Victor Kochkodan for their contribution to this research.
AbbrEvIAtIonS And SyMbolS<br />
NOM Natural organic matter<br />
qDMAEMA Quaternary 2-dimethyl-aminoethylmethacrylate<br />
PES Polyethersulphone<br />
PVDF Polyvinylidene fluoride<br />
HEMA Hydroxyethyl methacrylate<br />
DM Degree of modification<br />
d p Measured pore size (diameter) m<br />
� St<strong>and</strong>ard deviation of measurements<br />
� 0 Modal value of d p m<br />
PSD Pore size distribution<br />
BSA Bovine serum albumin<br />
SPEEK Sulphonated poly(ether-ether ketone)<br />
PEEK Poly(ether-ether ketone)<br />
HA Humic acid<br />
references<br />
ABBREvIATIONs AND syMBOLs 169<br />
[1] N. <strong>Hilal</strong>, V. Kochkodan, L. Al-Khatib, G. Busca, Characterization of molecularly<br />
composite membranes using an atomic force microscope, Surf. Interface Anal. 33<br />
(2002) 672–675.<br />
[2] M. Khayet, Membrane surface modification <strong>and</strong> characterization by X-ray photoelectron<br />
spectroscopy <strong>and</strong> contact angle measurements, Appl. Surf. Sci. 238 (1–4) (2004)<br />
269–272.<br />
[3] A.M. Mika, R.F. Childs, Acid/base properties of poly(4-vinylpyridine) anchored<br />
within microporous membranes, J. Memb. Sci. 152 (1) (1999) 129–140.<br />
[4] M. Nystrom, S. Butylin, S. Platt, NF retention <strong>and</strong> critical flux of small hydrophilic/<br />
hydrophobic molecules, Memb. Technol. 10 (2000) 1–5.<br />
[5] A.R. Costa, M.N. de Pinho, Effect of membrane pore size <strong>and</strong> solution chemistry on<br />
the ultrafiltration of humic substances solutions, J. Memb. Sci. 255 (1–2) (2005) 49–56.<br />
[6] L.-C. Xu, B.E. Logan, Interaction forces between colloids <strong>and</strong> protein-coated surfaces<br />
measured using an atomic force microscope, Environ. Sci. Technol. 39 (10) (2005)<br />
3592–3600.<br />
[7] L.-C. Xu, V. Vadillo-Rodriguez, B.E. Logan, Residence time, loading force, pH, <strong>and</strong><br />
ionic strength affect adhesion forces between colloids <strong>and</strong> biopolymer coated surfaces,<br />
Langmuir 21 (2005) 7491–7500.<br />
[8] N. <strong>Hilal</strong>, W.R. <strong>Bowen</strong>, L. Al-Khatib, O.O. Ogunbiyi, A review of atomic force microscopy<br />
applied to cell interactions with membranes, Chem. Eng. Res. Design 84 (A4)<br />
(2006) 282–292.<br />
[9] W.R. <strong>Bowen</strong>, N. <strong>Hilal</strong>, R.W. Lovitt, C.J. Wright, Direct measurement of the force<br />
of adhesion of a single biological cell using an atomic force microscope, Colloids Surf.<br />
A Physicochem. Eng. Asp. 136 (1998) 231–234.
170 5. AFM AND DEvELOPMENT OF (BIO)FOULINg-REsIsTANT MEMBRANEs<br />
[10] W.R. <strong>Bowen</strong>, N. <strong>Hilal</strong>, R.W. Lovitt, C.J. Wright, Direct measurement of interactions<br />
between adsorbed protein layers using an atomic force microscope, J. Colloid Interface<br />
Sci. 197 (1998) 348.<br />
[11] W.R. <strong>Bowen</strong>, N. <strong>Hilal</strong>, R.W. Lovitt, C.J. Wright, A new technique for membrane characterisation:<br />
direct measurement of the force of adhesion using an atomic force microscope,<br />
J. Memb. Sci. 139 (1998) 269–274.<br />
[12] W.R. <strong>Bowen</strong>, N. <strong>Hilal</strong>, R.W. Lovitt, C.J. Wright, Characterisation of membrane surfaces:<br />
direct measurement of biological adhesion using an atomic force microscope, J. Memb.<br />
Sci. 154 (2) (1999) 205–212.<br />
[13] W.R. <strong>Bowen</strong>, N. <strong>Hilal</strong>, R.W. Lovitt, C.J. Wright, An atomic force microscopy study of<br />
the adhesion of a silica sphere to a silica surface – effects of surface cleaning, Colloids<br />
Surf. A Physicochem. Eng. Asp. 157 (1999) 117–125.<br />
[14] G. Toikka, R.A. Hayes, J. Ralston, Adhesion of iron oxide to silica studied by atomic<br />
force microscopy, J. Colloid Interface Sci. 180 (1996) 239–248.<br />
[15] S. Veeramansunei, M.R. Yalamanchili, J.D. Miller, Measurement of interaction forces<br />
between silica <strong>and</strong> alpha-alumina by atomic force microscopy, J. Colloid Interface Sci.<br />
184 (1996) 594–600.<br />
[16] J.E. Kilduff, S. Mattaraj, J.P. Pieracci, G. Belfort, Photochemical modification of<br />
poly(ether sulfone) <strong>and</strong> sulfonated poly(sulfone) nanofiltration membranes for control<br />
of fouling by natural organic matter, Desalination 132 (1–3) (2000) 133–142.<br />
[17] M. Taniguchi, G. Belfort, Low protein fouling synthetic membranes by UV assisted<br />
surface grafting modification: varying monomer type, J. Memb. Sci. 231 (1–2) (2004)<br />
147–157.<br />
[18] M. Ulbricht, H. Matuschewski, A. Oechel, H.G. Hicke, Photo-induced graft polymerization<br />
surface modifications for the preparation of hydrophilic <strong>and</strong> low protein-adsorbing<br />
ultrafiltration membranes, J. Memb. Sci. 115 (1) (1996) 31–47.<br />
[19] L. Al-Khatib, Development of (bio)fouling resistant membranes for water treatment<br />
applications, in: School of Chemical, Environmental <strong>and</strong> Mining Engineering,<br />
University of Nottingham, 2006.<br />
[20] N. <strong>Hilal</strong>, L. Al-Khatib, H. Al-Zoubi, R. Nigmatullin, Atomic force microscopy study of<br />
membranes modified by surface grafting of cationic polyelectrolyte, Desalination 185<br />
(2005) 1025–1035.<br />
[21] D. Kuzmenko, E. Arkhangelsky, S. Belfer, V. Freger, V. Gitis, Chemical cleaning of UF<br />
membranes fouled by BSA, Desalination 179 (1–3) (2005) 323–333.<br />
[22] W.G. Liu, F. Li, X.D. Zhao, K.D. Yao, Q.G. Liu, Atomic force microscope characterisation<br />
forces between bovine serum albumin <strong>and</strong> cross-linked alkylated chitosan membranes in<br />
media of different pH, Polym. Int. 51 (12) (2002) 1459–1463.<br />
[23] A.I. Schafer, U. Schwicker, M.M. Fischer, A.G. Fane, T.D. Waite, Microfiltration of<br />
colloids <strong>and</strong> natural organic matter, J. Memb. Sci. 171 (2) (2000) 151–172.<br />
[24] J.A. Brant, Membrane colloid interactions: comparison of extended DLVO predictions<br />
with AFM force measurements, Environ. Eng. Sci. 19 (6) (2002) 413–427.<br />
[25] W.R. <strong>Bowen</strong>, T.A. Doneva, Atomic force microscopy of membranes, in: A. Hubbard<br />
(Ed.), Encyclopedia of Surface <strong>and</strong> Colloid Science, Marcel Dekker, New York, 2002.<br />
[26] A. Noshay, L.M. Robeson, Sulfonated polysulfone, J. Appl. Polym. Sci. 20 (1976)<br />
1885–1903.<br />
[27] M. Nystrom, P. Jarvinen, Modification of polysulfone ultrafiltration membranes with<br />
UV irradiation <strong>and</strong> hydrophilicity screening agents, J. Memb. Sci. 60 (1991) 275–296.<br />
[28] T. Kobayashi, M. Rikukawa, K. Sanui, N. Ogata, Proton-conducting polymers derived<br />
from poly(ether-ether ketone) <strong>and</strong> poly(4-phenoxybenzoyl-1,4-phenylene), Solid State<br />
Ionics 106 (1998) 219–225.
REFERENCEs 171<br />
[29] W.R. <strong>Bowen</strong>, T.A. Doneva, H.-B. Yin, Polysulfone – sulfonated poly(ethyl ether)<br />
ketone blend membranes: systematic synthesis <strong>and</strong> characterisation, J. Memb. Sci. 181<br />
(2001) 253–263.<br />
[30] W.R. <strong>Bowen</strong>, T.A. Doneva, H.-B. Yin, Separation of humic acid from a model surface<br />
water with PSU/SPEEK blend UF/NF membranes, J. Memb. Sci. 206 (2002) 417–429.<br />
[31] S.-H. Yoon, J.-S. Kim, Effect of the origin of humic acids on nanofilter fouling in drinking<br />
water production, Chem. Eng. Sci. 55 (2000) 5171–5175.
C H A P T E R<br />
6<br />
Nanoscale Analysis<br />
of Pharmaceuticals by<br />
Scanning Probe Microscopy<br />
Clive J. Roberts<br />
O U T L I N E<br />
6.1 Introduction 173<br />
6.2 The AFM as a Force Measurement Tool in Pharmaceuticals 174<br />
6.2.1 Particle Interaction Measurements 175<br />
6.2.2 Mechanical Properties from Single-particle Measurements 179<br />
6.3 AFM Imaging-Based Studies 183<br />
6.4 Micro- <strong>and</strong> Nanothermal Characterisation with SPM 185<br />
6.5 Conclusions 190<br />
Acknowledgements 190<br />
Abbreviations <strong>and</strong> Symbols 190<br />
References 191<br />
6.1 INTroduCTIoN<br />
There is an increasing dem<strong>and</strong> to develop pharmaceuticals <strong>and</strong> biomedical<br />
devices with architectures <strong>and</strong> complex chemistries controlled at<br />
the nanoscale. The need to develop delivery systems capable of targeting<br />
specific sites in the body or to successfully formulate poorly soluble<br />
drugs frequently requires nanoscale solutions. For example, targeted systems<br />
typically employ nanoparticles, incorporating a number of elements<br />
Atomic Force Microscopy in Process Engineering 173<br />
© 2009, Elsevier Ltd
174 6. NANOSCALE ANALySIS Of PHARMACEUTICALS by SCANNINg PRObE MICROSCOPy<br />
aimed at providing storage of the active therapeutic ‘stealth’ functionality<br />
to avoid the body defences as well as a targeting element. The additional<br />
trend towards the delivery of protein- <strong>and</strong> DNA-based medicines<br />
through new alternative delivery routes such as inhalation has also<br />
increased this need for nanoscale analysis. Such complex systems require<br />
analysis capable of minimal disruption due to sample preparation <strong>and</strong><br />
the ability to operate in a ‘physiological’ environment. This dem<strong>and</strong><br />
often exceeds the capacity of traditional analytical approaches to provide<br />
an effective underst<strong>and</strong>ing of this new generation of therapeutics.<br />
This chapter will highlight the role of atomic force microscopy (AFM)<br />
<strong>and</strong> other scanning probe microscopies (SPM) in the qualitative <strong>and</strong><br />
quantitative analysis of pharmaceuticals, highlighting both imaging <strong>and</strong><br />
force data acquisition modes. These microscopes have enabled the study<br />
of pharmaceutical devices [1–3] <strong>and</strong> drug particles [4–6] with minimal<br />
pre-treatment in both air <strong>and</strong> liquid at the nanoscale level. The potential<br />
for SPMs, <strong>and</strong> AFM in particular, is now being exploited as an integral<br />
part of formulation, in both academic <strong>and</strong> industrial research. It is of<br />
course important to note that whilst SPM-based solutions do represent an<br />
increasingly important part of the analysis of pharmaceuticals, they are<br />
best applied with complementary approaches, such as Raman spectroscopy<br />
or diffraction-based techniques, which can provide chemically specific<br />
<strong>and</strong> 3D structural data.<br />
6.2 The AFM AS A ForCe MeASureMeNT Tool IN<br />
PhArMACeuTICAlS<br />
Most pharmaceuticals are produced as solid dosage forms (e.g. tablets,<br />
capsules, inhalable powders) <strong>and</strong> hence their manufacture inevitably<br />
involves the h<strong>and</strong>ling <strong>and</strong> manipulation of powdered materials. These<br />
powders can have particles ranging from submicron to many hundreds<br />
of microns in size. An underst<strong>and</strong>ing of how the mechanical properties<br />
of these particles <strong>and</strong> the forces between them influence factors such as<br />
processability <strong>and</strong> stability is an important feature of formulation development.<br />
Indeed, in some cases, the final form of the drug (<strong>and</strong> other<br />
materials in the medicine (the excipients)) is loose powder, as e.g. in a dry<br />
powder inhaler (DPI). Here then, the successful delivery of the drug itself<br />
relies on an intimate underst<strong>and</strong>ing of particle properties <strong>and</strong> their interactions.<br />
Until recently, the direct assessment of particle-particle <strong>and</strong> particle-device<br />
interactions relied upon long-established bulk methods that<br />
deal with large numbers of particles, such as the centrifugal technique<br />
[7, 8]. Whilst effects such as cohesion <strong>and</strong> adhesion phenomena can be<br />
studied, these data reveal little concerning the nature <strong>and</strong> interplay of the
6.2 THE AfM AS A fORCE MEASUREMENT TOOL IN PHARMACEUTICALS 175<br />
fundamental forces involved (e.g. van der Waals, electrostatic, capillary).<br />
Access to such information is beneficial not only to the assessment of a<br />
formulation, but also to establish the basis of any required particle modification<br />
<strong>and</strong> optimisation.<br />
Just as particle interactions can be probed using AFM, the nanoscale<br />
mechanical properties of powders can also be investigated. Previously,<br />
this could only be derived from bulk techniques, where powders are<br />
compressed into miniature beams <strong>and</strong> three-point bending tests are<br />
performed. The need for relatively large amounts of powder <strong>and</strong> to<br />
account for the porosity of the beams has limited the applicability of this<br />
approach. Even using miniaturised beams, milligram quantities of material<br />
are required, preventing screening of active pharmaceutical ingredients<br />
at early stages of development [9, 10]. Here, we consider how AFM<br />
has been used to address both particle interactions <strong>and</strong> the measurement<br />
of the mechanical properties from single particles.<br />
6.2.1 Particle Interaction Measurements<br />
The ability to study single-particle interactions <strong>and</strong> the forces involved<br />
became possible with the advent of the AFM in 1986 [11–13]. In particular,<br />
the use of AFM in the so-called ‘colloidal probe’ technique, whereby<br />
the force of interaction between a spherical bead attached to the AFM<br />
cantilever <strong>and</strong> a planar surface was studied, revealed the potential of this<br />
approach [14]. Importantly, a single particle (e.g. drug) attached to the<br />
AFM cantilever can be used for a series of comparative experiments challenging<br />
different substrates. In addition, the ability of AFM to work in a<br />
variety of environments, such as controlled humidity <strong>and</strong> in liquids, is<br />
significant for pharmaceutical applications.<br />
The first example of this approach being used for a pharmaceutical<br />
powder examined the differences in the adhesion of lactose particles to<br />
two gelatin DPI capsule surfaces [15]. It is important when attaching<br />
such individual drug particles to an AFM cantilever that their contacting<br />
region remains free from any adhesives employed or damage during<br />
attachment. An example scanning electron microscope (SEM) image is<br />
shown in Figure 6.1, where a drug particle is attached to an AFM cantilever.<br />
Typically, to prepare such a probe would involve the use of a minute<br />
amount of glue on the end of the cantilever to which a particle is attached<br />
using a micromanipulator or the AFM itself. This relatively labourintensive<br />
sample preparation currently limits the number of different<br />
particles used within one study (usually to no more than five). The accessible<br />
particle size is typically in the range between 0.5 �m <strong>and</strong> 50 �m. For<br />
very small <strong>and</strong>/or cohesive particles (1 �m or less), more than one may<br />
become attached to the lever; however, as long as only one comes into<br />
contact with the surface to be challenged, this is acceptable.
176 6. NANOSCALE ANALySIS Of PHARMACEUTICALS by SCANNINg PRObE MICROSCOPy<br />
500 nm<br />
AFM cantilever<br />
Drug particle<br />
FIgure 6.1 SEM image of the end of an AFM cantilever to which a single drug particle<br />
has been attached. The lower extremity of this particle forms the contact region with a sample<br />
surface being studied within the AFM.<br />
Having manufactured a particle probe, the forces of interaction of this<br />
probe with a surface are recorded. Figure 6.2 illustrates a schematic of a<br />
force curve <strong>and</strong> the main stages in acquiring such single-particle adhesion<br />
data. If such data is to be quantified in terms of force, then the<br />
spring constant of the AFM lever needs to be determined to enable calibration<br />
[16]. To ease comparisons between different-particle adhesion<br />
measurements, it is important to control key environmental factors such<br />
as humidity <strong>and</strong> to keep such instrument parameters as maximum presson<br />
force, contact time, particle approach <strong>and</strong> retract speed constant.<br />
Many researchers have now exploited this approach to explore particle<br />
interactions between drug <strong>and</strong> excipient particles <strong>and</strong> the components<br />
of delivery devices. Louey et al. used this approach to measure pulloff<br />
forces between a model colloidal silica probe <strong>and</strong> lactose particles<br />
suitable for use as carriers in a DPI [17]. AFM has also been used to<br />
rank the force of interaction of salbutamol with materials relevant to its<br />
inhaled delivery in the following manner, glass � lactose � salbutamol �<br />
polytetrafluoroethylene (PTFE), showing that on PTFE tribocharging<br />
occurred following repeated contact [18]. Using AFM in this way has<br />
also revealed the effect of inducing amorphous content on the surface of<br />
particles of the drug zanamivir, where an increase in its affinity with a lactose<br />
carrier surface was observed [19]. Such surface amorphous material is<br />
important from a pharmaceutical perspective as this will be more soluble<br />
<strong>and</strong> probably less stable than a crystalline form of the drug, <strong>and</strong> hence can<br />
cause desirable (or undesirable) effects in terms of increased amounts of<br />
drug delivered on administration <strong>and</strong> stability problems in a medicine.
6.2 THE AfM AS A fORCE MEASUREMENT TOOL IN PHARMACEUTICALS 177<br />
(a)<br />
Cantilever deflection (nm)<br />
(b)<br />
Motion of<br />
lever<br />
0<br />
AFM cantilever<br />
with drug particle<br />
Sample<br />
Particle approaching sample<br />
Particle adhered<br />
to sample<br />
Particle detached<br />
to sample<br />
0<br />
Relative veticle position of sample <strong>and</strong> particle (nm)<br />
FIgure 6.2 (a) Schematic illustrating an AFM probe with attached drug particle coming<br />
to, contacting with <strong>and</strong> being removed from a sample surface. (b) With data recorded<br />
from such an event, the cantilever deflection can be converted to a force of the spring constant<br />
if the lever is known <strong>and</strong> the relative positions to a true sample-particle separation if<br />
the sensitivity of the microscope has been calibrated.<br />
The influence of relative humidity on the forces between particles is<br />
an important consideration in pharmaceuticals as this can strongly influence<br />
the stability of a medicine through the mediation of solution-based<br />
processes owing to surface adsorbed water. In addition, owing to the formation<br />
of capillary bridges between particles with increased humidity,<br />
this may cause aggregation. Hence, the effect of humidity has been the<br />
subject of a number of AFM studies. Young <strong>and</strong> co-workers have shown<br />
that drug cohesion increases at elevated humidity for some materials but<br />
decreased for others, potentially as a result of long-range attractive electrostatic<br />
interactions [20]. A variation in particle-contact morphology has<br />
also been shown to cause similar behaviour with increasing humidity<br />
[21]. The ability to record force data in controlled environments has been<br />
extended to work in liquids, e.g. to rank interactions in model propellants<br />
so as to simulate the environment within a pressurised metered dose<br />
inhaler [22–24].
178 6. NANOSCALE ANALySIS Of PHARMACEUTICALS by SCANNINg PRObE MICROSCOPy<br />
These <strong>and</strong> other studies can broadly be divided into two types: those<br />
that rank relative particulate interactions <strong>and</strong> those that attempt to make<br />
a quantitative comparison of force per unit area of contact. Ranking studies<br />
address, e.g., how drug-drug cohesion compares to drug-excipient<br />
particle <strong>and</strong> drug-device adhesion. Since particulate interactions are dominated<br />
by aspects such as surface morphology, surface roughness, exposed<br />
chemical moieties <strong>and</strong> thermodynamic properties, all of which can vary<br />
from particle to particle <strong>and</strong> indeed within a single particle, such ranking<br />
comparisons are normally made using the same particle to challenge<br />
all the possible combinations of interactions. In this case, ranking should<br />
be consistent for each drug particle probe, but the absolute values of adhesion<br />
force determined cannot be used to make comparisons from particle<br />
to particle or between materials [18].<br />
To quantify such measurements, particle variability <strong>and</strong> a lack of direct<br />
knowledge of the contacting regions must be overcome to allow the<br />
determination of factors such as surface energy <strong>and</strong> work of adhesion.<br />
The use of AFM to determine such properties on model flat surfaces, such<br />
as silicon, was established relatively early [25]. However, its application<br />
to pharmaceuticals due to difficulties of rough <strong>and</strong> variable particle morphology<br />
came later [19, 21, 22]. In these works, the principal variable that<br />
is allowed for is contact area. Contact areas have, to date, been estimated<br />
either via a direct imaging approach [22] or indirectly through imaging<br />
indents made by the particle probe in a plastic polymer film [19]. The<br />
subsequent use of adhesion models such as Johnson–Kendall–Roberts<br />
<strong>and</strong> Derjaguin–Muller–Toporov can then allow the surface energy of the<br />
particle (over the region of contact) to be determined. In this way, e.g.,<br />
micronised (milled) salbutamol sulphate <strong>and</strong> a version prepared via a<br />
novel supercritical fluid method have been compared [22].<br />
An alternative to this approach of trying to model the variable contact<br />
region between particles is to compare ratios of cohesive <strong>and</strong> adhesive<br />
forces between different particles rather than actual separation forces.<br />
This has allowed an assessment for relatively flat crystals of the affinity<br />
of salbutamol sulphate to lactose <strong>and</strong> budesonide to lactose, showing<br />
that salbutamol sulphate has a stronger affinity for lactose [26]. This<br />
information was then related successfully to the likely blend uniformity<br />
these materials would form.<br />
In addition to monitoring force normal to a surface, AFM is also capable<br />
of assessing frictional forces between a probe <strong>and</strong> a surface. This is<br />
achieved by recording the twist of the cantilever in addition to its vertical<br />
bend as it scans a surface in continuous contact with that surface<br />
(Figure 6.3a). This approach has recently been used to obtain singleparticle<br />
friction measurements on DPI formulations [27] <strong>and</strong> blister<br />
packaging material (used in DPIs) [28] <strong>and</strong> provides an opportunity to<br />
consider sliding as well as separation forces. Figure 6.3 shows examples
6.2 THE AfM AS A fORCE MEASUREMENT TOOL IN PHARMACEUTICALS 179<br />
(a) (b)<br />
Scan<br />
direction<br />
(c)<br />
Position 1<br />
Lateral<br />
force<br />
signal<br />
Position 2<br />
Particle A<br />
Particle B<br />
Particle C<br />
PTFE Glass Zanamivir Zanamivir Magnesium<br />
�<br />
Magnesium<br />
stearate<br />
stearate<br />
PTFE<br />
of the results of recording the sliding friction of single lactose particles<br />
on various drugs, excipients <strong>and</strong> materials used in inhaler devices. This<br />
showed a ranking of glass�zanamivir � zanamivir–magnesium stearate<br />
blend�magnesium stearate�PTFE (see Figure 6.3c). The addition of magnesium<br />
stearate (a commonly employed lubricant) to the drug zanamivir<br />
was seen to dominate the behaviour of this system <strong>and</strong> significantly<br />
reduced the friction [27].<br />
6.2.2 Mechanical Properties from Single-particle Measurements<br />
Consideration of the schematic force distance curve in Figure 6.2(b)<br />
reveals that not only can adhesion data be determined, but if the region<br />
of contact is considered where the probe is pressed into a surface <strong>and</strong><br />
is then withdrawn, then it is clear that this contains information on<br />
the mechanical properties of the sample. Indeed, data gathered from a<br />
nanoscale contact in this manner is similar to traditional indentation<br />
Drug<br />
MgSt<br />
Glass<br />
Drug �<br />
MagSt<br />
FIgure 6.3 (a) Schematic of AFM particle friction set-up. Position 1: low-friction surface<br />
causes very small lateral force on cantilever. Position 2: high-friction surface causes<br />
large lateral force <strong>and</strong> cantilever twists. Inset: SEM image of lactose particle attached to cantilever.<br />
(b) 5 �m � 5 �m topographic AFM images of surfaces under study. From the top,<br />
clockwise: PTFE; glass; the drug zanamivir with magnesium stearate; magnesium stearate<br />
<strong>and</strong> zanamivir. (c) Processed data showing friction on all surfaces for three lactose particles.
180 6. NANOSCALE ANALySIS Of PHARMACEUTICALS by SCANNINg PRObE MICROSCOPy<br />
measurements more typically measured on a micron or greater scale to<br />
determine elastic modulus <strong>and</strong> hardness [29]. There are, however, a number<br />
of limitations of the AFM data that must first be considered. First,<br />
as st<strong>and</strong>ard AFM cantilevers are typically very flexible (so that they are<br />
sensitive to nanoNewton forces), they are incapable of deforming surfaces<br />
beyond approximately 10 GPa elastic modulus. Second, unlike a<br />
traditional indenter, an AFM probe does not approach or deform a surface<br />
completely normal to that surface, because of the need to have the<br />
lever at a slight angle to ensure that the probe apex contacts the sample<br />
first. This causes lateral deformation errors in the data, which are only<br />
negligible with relatively small indent depths.<br />
Despite these issues, many groups have employed AFM to determine the<br />
elastic, inelastic <strong>and</strong> hardness properties of materials. Such deformation<br />
behaviour of pharmaceutical ingredients is known to affect pharmaceutical<br />
processes such as milling <strong>and</strong> compaction [30–33]. To date, most<br />
reported pharmaceutical examples of this approach have used a modified<br />
AFM equipped with a relatively large probe with well-defined<br />
geometry <strong>and</strong> a much stiffened spring that supports this probe [30,<br />
34, 35]. Typically, in these methods, material parameters are extracted<br />
from load-displacement unloading curves using approaches outlined by<br />
Oliver <strong>and</strong> Pharr [36]. Recently, Ward <strong>and</strong> co-workers [37] have utilised<br />
a sharp AFM probe <strong>and</strong> normal cantilevers to record nanomechanical<br />
measurements on sorbital samples to quantitatively distinguish between<br />
amorphous <strong>and</strong> crystalline domains through Young’s modulus measurements.<br />
Figure 6.4 shows images <strong>and</strong> typical force distance curves<br />
obtained from crystalline <strong>and</strong> amorphous sorbitol regions <strong>and</strong> a control<br />
of a hard, non-indenting silicon substrate. To determine the nanoindentation<br />
of the probe tip into the sorbitol sample as a function of load,<br />
the force distance curves from the sorbitol sample <strong>and</strong> the control hard<br />
substrate were compared. When comparing the amorphous <strong>and</strong> the<br />
crystalline regions, it is the amorphous region that provided the greater<br />
deflection of the tip (indicating the softer material <strong>and</strong> greater indentation)<br />
at the given applied loads [37]. In this approach, a comparison of<br />
the gradients of the contact region of the force distance curves between<br />
a hard non-deformable reference surface <strong>and</strong> the sample can provide the<br />
relative stiffness of a surface [38]. By applying the Hertz model [39] to<br />
the nanoindentation data, a quantitative value of the Young’s modulus of<br />
the surfaces can be obtained. The Hertz model describes the elastic deformation<br />
of two homogenous surfaces under an applied load <strong>and</strong> is often<br />
used to model AFM data since it requires little knowledge of parameters<br />
such as surface energy. The first step in modelling the data is to compare<br />
force distance curves recorded on the sample <strong>and</strong> an ideally hard reference<br />
(glass surface) to determine the indentation (�) of the probe into the<br />
sample as a function of load. The contact regions of the curves are overlaid<br />
such that zero force is equal to zero indentation.
(a)<br />
4 µm<br />
0<br />
(b)<br />
4 µm<br />
0<br />
6.2 THE AfM AS A fORCE MEASUREMENT TOOL IN PHARMACEUTICALS 181<br />
Force (nN)<br />
(e)<br />
8<br />
7<br />
6<br />
5<br />
4<br />
3<br />
2<br />
1<br />
0<br />
�1<br />
Retracting<br />
0 20 40 60<br />
Log E (Pa)<br />
Force (nN)<br />
The value of indentation can then be related to the combined elastic<br />
modulus of the tip <strong>and</strong> the sample (K) by:<br />
L<br />
K �<br />
( � R)<br />
10<br />
9<br />
8<br />
7<br />
6<br />
5<br />
4<br />
3<br />
2<br />
1<br />
3 2<br />
0 50 100 150<br />
Relative z (nm)<br />
200 250 300<br />
1<br />
1. Si Reference<br />
2. Crystalline<br />
3. Amorphous<br />
1.0 � 10 8<br />
1.0 � 10 7<br />
1.0 � 10 6<br />
1.0 � 10 5<br />
1.0 � 10 4<br />
2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0<br />
Load (nN)<br />
Extending<br />
80 100 120 140 160 180<br />
Relative z (nm)<br />
3 1/ 2<br />
8<br />
Crystalline<br />
Amorphous<br />
FIgure 6.4 (a) AFM images of unmodified sorbitol (topographic height <strong>and</strong> phase<br />
images), insert highlights b<strong>and</strong>ing pattern; (b) AFM images within a quench-cooled<br />
domain of sorbitol, which would be expected to have a more amorphous nature;<br />
(c) nanoindentation curves for the crystalline <strong>and</strong> amorphous domains; (d) load dependence<br />
on Young’s modulus <strong>and</strong> (e) force curve from the amorphous domain. Reproduced<br />
with permission [37].<br />
(6.1)<br />
where L is the load <strong>and</strong> R is the radius of the probe. Since the combined<br />
elastic modulus contains the elastic moduli for the tip <strong>and</strong> the sample,<br />
(c)<br />
(d)
182 6. NANOSCALE ANALySIS Of PHARMACEUTICALS by SCANNINg PRObE MICROSCOPy<br />
an expression containing the E s, the elastic moduli of the sample can be<br />
obtained:<br />
� t � s<br />
K � �<br />
E E<br />
�<br />
4⎛<br />
2 2<br />
1 υ 1 υ ⎞<br />
⎜<br />
3 ⎝⎜<br />
⎠⎟<br />
t<br />
s<br />
1<br />
(6.2)<br />
where ν t <strong>and</strong> ν s are the Poisson’s ratio of the tip <strong>and</strong> the sample, respectively.<br />
Since E t is much greater than E s, the first term in the bracket of<br />
equation (6.2) may be ignored to produce equation (6.3):<br />
E<br />
s<br />
3K( 1 � υs)<br />
�<br />
4<br />
Combining equations (6.1) <strong>and</strong> (6.3) derives:<br />
E<br />
s<br />
2<br />
2<br />
3L( 1 � υs<br />
)<br />
� 3 1/ 2<br />
4( 8�<br />
R)<br />
(6.3)<br />
(6.4)<br />
The value of ν s is difficult to determine accurately for many samples<br />
such as sorbitol <strong>and</strong> other pharmaceutical materials. Therefore, it is<br />
appropriate to use a midrange value of ν s (typical range is from 0.1–0.5)<br />
since varying this has little effect on the value obtained for E s.<br />
The Hertz model is valid only for elastic deformation, so it is important<br />
to remove any contribution from inelastic deformation from the analysis.<br />
The model also assumes no significant adhesion between the probe <strong>and</strong><br />
sample. To overcome this, only data from the loading curve is typically<br />
used. To assess the validity of the data for use with the Hertz model, a plot<br />
of load versus indentation is typically produced. In the ideal situation, a linear<br />
relationship between load <strong>and</strong> indentation should be observed to allow<br />
E s to be calculated. The Young’s modulus values for the amorphous regions<br />
of the sorbitol using this type of analysis of Ward et al. were less than the<br />
crystalline regions [37]. Interestingly, the Young’s modulus values appeared<br />
to rise with increases in loading in both amorphous <strong>and</strong> crystalline regions<br />
of sorbital, suggesting possible viscoelastic/plastic deformation [37]. Hence,<br />
these surface mechanical measurements were able to distinguish between<br />
crystalline <strong>and</strong> amorphous material <strong>and</strong> exemplify the possible use of the<br />
AFM to screen batch-to-batch variations in drug formulations.<br />
In general, it can be seen, therefore, that AFM provides a very flexible<br />
platform, adaptable to a range of experimental geometries for force measurements<br />
with real pharmaceutical materials. Examples of determining<br />
the interaction between an iron-coated AFM probe <strong>and</strong> a range of drugs<br />
to simulate tablet press–tablet interactions [40], <strong>and</strong> particle-bubble interaction<br />
forces in the mineral processing industry [41] give an idea of other<br />
possibilities yet to be fully explored.
6.3 AfM IMAgINg-bASEd STUdIES 183<br />
6.3 AFM IMAgINg-BASed STudIeS<br />
As a microscopy, AFM is perhaps better known for its ability to image<br />
surfaces at nanometre resolution in a variety of environments than as a<br />
force measurement tool. As a material under study can be exposed to<br />
environmental stresses (temperature, humidity, solvents etc.), it can often<br />
be in a dynamic state <strong>and</strong> hence, as long as the kinetics are not too rapid,<br />
AFM is able to provide a dynamic view of surface-mediated processes<br />
(unless employing advanced video rate imaging AFM [42], a typical<br />
1 �m � 1 �m image takes approximately 30 s to acquire). This approach<br />
has found effective applications in pharmaceuticals in the study of environmental<br />
stress on the surface properties of formulations, drug release<br />
from erodible materials <strong>and</strong> crystallisation phenomena.<br />
Price <strong>and</strong> Young [43] have studied the real-time effects of changes in<br />
relative humidity (RH) on spray-dried lactose using environmentally<br />
controlled AFM. The lactose was imaged at 0%, 30% <strong>and</strong> 58% RH for 4, 2<br />
<strong>and</strong> 22 h, respectively. Recrystallisation of amorphous lactose, promoted<br />
by the presence of water, was observed at 58% <strong>and</strong> 75% RH, although<br />
only some of the particles were shown to undergo nucleation <strong>and</strong> crystal<br />
growth. The AFM data was combined with traditional approaches<br />
such as X-ray powder diffraction, differential scanning calorimetry <strong>and</strong><br />
dynamic vapour sorption <strong>and</strong> was shown to have significant potential<br />
for studies of the nucleation <strong>and</strong> crystallisation processes of amorphous<br />
material [43].<br />
Li <strong>and</strong> co-workers have utilised contact mode AFM imaging in air to<br />
image a series of partial dissolutions on the (010) face of paracetamol<br />
[44]. From the AFM images of the (010) face of acetaminophen crystals<br />
treated with alternative dissolution solvents, different etching patterns<br />
were observed. These different etching patterns were speculated to be a<br />
result from the surface diffusion of the crystal molecules desorbed during<br />
the dissolution process <strong>and</strong> by the mutual interaction between the<br />
solvent <strong>and</strong> crystal molecules. AFM has similarly been employed to<br />
visualise the effect of dissolution media on various polymeric systems<br />
designed to achieve a specific drug-release profile [45].<br />
It is widely known that pharmaceutical additives <strong>and</strong> impurities in<br />
drug formulations can affect drug crystal growth, morphology <strong>and</strong> potential<br />
drug efficacy. This may be achieved deliberately to control crystal<br />
habit or an undesirable consequence of contamination. By underst<strong>and</strong>ing<br />
the degree to which impurities affect these critical formulation parameters,<br />
drug delivery systems can be developed more effectively. AFM has<br />
the capability to study drug crystal changes in real time, as demonstrated<br />
by Thompson et al. in the study of paracetamol in the presence of impurities,<br />
acetanilide <strong>and</strong> metacetamol using AFM in liquid [46]. A series of<br />
real-time AFM images of the (001) face of a native paracetamol crystal are
184 6. NANOSCALE ANALySIS Of PHARMACEUTICALS by SCANNINg PRObE MICROSCOPy<br />
FIgure 6.5 A sequence of 10 �m � 10 �m AFM images of the (001) face of a paracetamol<br />
crystal during incubation in paracetamol solution. In image (a) time t � 0 s, (b) t � 60 s,<br />
(c) t � 118 s, (d) t � 176 s, (e) t � 234 s <strong>and</strong> (f) t � 292 s. Figure reproduced with permission [46].<br />
shown in Figure 6.5, <strong>and</strong> illustrates that the growth steps of crystallisation<br />
over 4.8 h are curved in appearance, indicative of crystal growth via<br />
the screw dislocation mechanism. AFM images of the (001) face of another<br />
paracetamol crystal during incubation with 4 mol% acetanilide over time<br />
are shown in Figure 6.6. The defects in the crystal surface form weak<br />
points, leading to deepened holes (one hole is highlighted by a black<br />
arrow in Figure 6.6) over 12 h, indicating that the dissolution occurs both<br />
laterally <strong>and</strong> towards the crystal core [46]. Images during incubation with<br />
4 mol% metacetamol showed that the steps formed were significantly<br />
smaller <strong>and</strong> pointed in appearance (image not shown here). These features<br />
were ascribed to ‘pinning’, an effect caused when additives adsorb<br />
onto the steps <strong>and</strong> force them to grow between the impurities [46].<br />
From our consideration of AFM as a force measurement tool, it is clear<br />
that the probe is sensitive not only to topography but also to the magnitude<br />
<strong>and</strong> nature of the forces experienced by the probe. Consequently,<br />
whilst imaging, if such sensitivity can be mapped also, it becomes possible<br />
to produce alongside topography data sensitive to a particular surface<br />
property (e.g. adhesion, harness etc.). For example, during the commonly<br />
employed tapping mode imaging [47], where the AFM cantilever is oscillated<br />
at its resonant frequency <strong>and</strong> intermittently contacts a sample surface<br />
at the bottom of each oscillation cycle, the phase of the oscillation as well as
6.4 MICRO- ANd NANOTHERMAL CHARACTERISATION wITH SPM 185<br />
FIgure 6.6 A sequence of 10 �m � 10 �m AFM images of the (001) face of a paracetamol<br />
crystal in paracetamol/4 mol% acetanilide solution. The dissolution of step one<br />
is followed by a black arrow on each image. Image (a) is taken 11.5 min after addition of<br />
paracetamol/4 mol% acetanilide solution. Images (b–f) are taken at the following times<br />
after image (a): (b) t � 146 s, (c) t � 290 s, (d) t � 436 s, (e) t � 582 s <strong>and</strong> (f) t � 728 s. Figure<br />
reproduced with permission [46].<br />
the topography can be recorded. Phase images are obtained by measuring<br />
the phase lag between the signal that drives the cantilever-tip oscillation<br />
<strong>and</strong> the output signal of the cantilever oscillation (will differ more to the<br />
original cantilever-tip oscillation if the sample surface is soft <strong>and</strong> viscoelastic<br />
due to greater energy lost to the sample). The use of phase imaging<br />
alongside topographical imaging is particularly useful for studying drug<br />
polymorphs phase-separated systems <strong>and</strong> drug-excipients systems as it<br />
can be used to map out a profile of materials with different mechanical <strong>and</strong><br />
physicochemical properties [48, 49].<br />
6.4 MICro- ANd NANoTherMAl<br />
ChArACTerISATIoN wITh SPM<br />
Knowledge of the distribution <strong>and</strong> physical form of a drug in its<br />
formulation is important in determining its overall performance <strong>and</strong><br />
can further our underst<strong>and</strong>ing into bioavailability issues in medicines.<br />
Classically, one of the main tools used to identify the form a drug takes<br />
within a formulation <strong>and</strong> the interactions between its components is
186 6. NANOSCALE ANALySIS Of PHARMACEUTICALS by SCANNINg PRObE MICROSCOPy<br />
calorimetry. However, such approaches cannot reveal the spatial distribution<br />
of these components. An adaptation of AFM that can achieve this is<br />
the Scanning Thermal Microscope (SThM) [50]. Here, the normal silicon<br />
cantilever is replaced with a Wollaston wire probe, brought to a sharp<br />
apex in the form of a cantilever. The spatial (<strong>and</strong> thermal mapping) resolution<br />
of such probes is on the order of a micron. Whilst this is useful,<br />
they also have the additional ability to apply heat locally to a surface <strong>and</strong><br />
to record the power required to deliver a constant temperature increase to<br />
that point in the sample (in a manner analogous to Differential Scanning<br />
Calorimetry (DSC)), therefore providing detailed thermal characterisation<br />
at that point on a sample (the so-called Local Thermal Analysis (LTA)).<br />
To date, LTA has been the main area of application of this approach in<br />
pharmaceuticals.<br />
Six et al. have demonstrated the ability of SThM analysis to distinguish<br />
<strong>and</strong> identify phase-separated material in solid dispersions of<br />
itraconazole <strong>and</strong> Eudragit ® E100 polymer [51]. Such dispersions of<br />
a poorly soluble drug in a soluble matrix represent a classic route to<br />
improving the bioavailability of such drugs. SThM images of dispersions<br />
containing 10%, 40% <strong>and</strong> 60% itraconazole showed an increase in<br />
surface roughness with an increase in drug loading, possibly linking the<br />
rough surface areas to phase-separated drug. Phase-separated regions of<br />
the drug are not desired as they would be less liable to release the active<br />
ingredient than molecular dispersed drug. LTA has been used to characterise<br />
the different phases by a comparison of the glass transition (T g) of<br />
different dispersions with that of the pure compounds of drug <strong>and</strong> polymer.<br />
LTA showed that the penetration of the probe tip into a sample of<br />
pure itraconazole occurred at around 333 K, which is in good agreement<br />
with the previously determined T g for the drug by DSC of 332.4 K. In contrast,<br />
the penetration of the probe tip into pure Eudragit ® E100 occurred<br />
at a much higher temperature of 383 K compared to its T g of 318 K determined<br />
by DSC <strong>and</strong> was linked to the fragility of the sample [51]. It was<br />
also observed in dispersions with low drug loading (10%) that the probetip<br />
penetration was at a temperature in between the T gs for itraconazole<br />
<strong>and</strong> Eudragit ® E100, an indication of intimate mixing of the components.<br />
At high drug loading (60%) dispersions, the T g was similar to that for the<br />
drug, indicating separation of the components.<br />
In another study by S<strong>and</strong>ers et al. [52], SThM combined with LTA<br />
was used to distinguish between two polymorphs of cimetidine, types<br />
A <strong>and</strong> B. These two pharmaceutically useful anhydrous polymorphs of<br />
cimetidine had previously been tentatively distinguished by AFM phase<br />
imaging in a variable humidity environment [48]. An underst<strong>and</strong>ing<br />
<strong>and</strong> ability to detect <strong>and</strong> distinguish between different polymorphs in<br />
drug formulations is of particular significance, as different polymorphs<br />
can have different physicochemical properties <strong>and</strong> can convert from one
6.4 MICRO- ANd NANOTHERMAL CHARACTERISATION wITH SPM 187<br />
form to another under storage conditions. While cimetidine is known to<br />
have seven polymorphs, only type A <strong>and</strong> type B are used in tablet <strong>and</strong><br />
suspension formulations, respectively [53]. SThM images were obtained<br />
using a probe temperature of 50°C <strong>and</strong> a scan rate of 1 Hz. Figure 6.7<br />
shows topographical (images a <strong>and</strong> c) <strong>and</strong> SThM images (images b <strong>and</strong> d)<br />
of a 50:50 mixture of polymorph type A <strong>and</strong> type B from two different<br />
areas. The contrast highlighted by arrows in images b <strong>and</strong> d is a result of<br />
the differences in surface thermal conductivity in the two different polymorphs.<br />
LTA on discs of pure samples of polymorphs type A <strong>and</strong> type B<br />
were used to help elucidate that the lighter <strong>and</strong> darker regions observed<br />
in the topographical data (Figure 6.7a, c) were of polymorphs type A <strong>and</strong><br />
FIgure 6.7 50 �m � 50 �m images of a pressed disc comprising a 50:50 mixture of the A<br />
<strong>and</strong> B polymorphs of cimetidine (scale bar: 10 mm): (a) <strong>and</strong> (c) show topographical images of<br />
two different areas of the disc <strong>and</strong> (b) <strong>and</strong> (d) show the corresponding thermal conductivity<br />
images. Features highlighted with arrows show contrast as a result of the differing thermal<br />
behaviour of the two polymorphic forms of cimetidine. Reproduced with permission [52].
188 6. NANOSCALE ANALySIS Of PHARMACEUTICALS by SCANNINg PRObE MICROSCOPy<br />
type B, respectively. LTA was performed with a temperature ramp rate<br />
of 10°C s �1 . Figure 6.8 shows the local thermal analysis of polymorphs B<br />
<strong>and</strong> A with melting points observed between 140–146°C <strong>and</strong> 141–143°C,<br />
respectively. The close range of melting points for polymorphs A <strong>and</strong> B<br />
illustrates why bulk thermal analytical methods have difficulty distinguishing<br />
between the two polymorphs. An interesting feature of the data<br />
for polymorph A is the endothermic peak seen just above 100°C (marked<br />
X in Figure 6.8b). This feature is seen in neither the bulk thermal analysis<br />
of form A nor LTA of form B. This would suggest that this feature is a<br />
result of the increased surface sensitivity of the localized measurements,<br />
<strong>and</strong> that this endothermic peak is likely to be a result of adsorbed surface<br />
water. This is consistent with form A being more hydrophillic than form<br />
(a)<br />
Sensor movement (μm) Sensor movement (μm)<br />
(b)<br />
1.5<br />
0<br />
–1.5<br />
–3<br />
1<br />
0<br />
–1<br />
–2<br />
(I)<br />
(III)<br />
20 60 100 140<br />
Temperature (°C)<br />
(I)<br />
(III)<br />
20<br />
60<br />
100 180<br />
Temperature (°C)<br />
FIgure 6.8 Localised thermal analysis of pure cimetidine polymorphs (a) B <strong>and</strong> (b) A.<br />
Lines denoted (I) show the calorimetric data for a single point <strong>and</strong> dashed lines (II) show<br />
the movement of the sensor during the measurement, corresponding to thermomechanical<br />
analysis. Lines (III) <strong>and</strong> (IV) show similar data for other points on the samples. Reproduced<br />
with permission [52].<br />
(II)<br />
(IV)<br />
(IV)<br />
(II)<br />
1<br />
0<br />
1<br />
0<br />
–2<br />
–4<br />
–2<br />
–2<br />
Derivative power (μW deg –1 )<br />
Derivative power (μW deg –1 )
6.4 MICRO- ANd NANOTHERMAL CHARACTERISATION wITH SPM 189<br />
B <strong>and</strong> with previous AFM data, which displayed a behaviour indicative<br />
of adsorbed water at the surface [48].<br />
These <strong>and</strong> other results not only demonstrate the ability of LTA to thermally<br />
characterise individual components within formulation mixes, but<br />
also highlight the limitation of the system to detect structures of approximately<br />
20 �m � 20 �m or larger, such that if more than one structure<br />
is present in a region, the LTA measurement appears as an intermediate<br />
softening of values between individual components.<br />
Clearly, replacement of the traditional AFM imaging probe with a<br />
heat-conductive Wollaston wire has enabled the local thermal properties<br />
of pharmaceuticals to be investigated but at the sacrifice of spatial resolution.<br />
However, the recent development of a commercially fabricated<br />
doped-silicon cantilever with a heated probe has enabled this problem<br />
to be overcome <strong>and</strong> for nanoscale thermal events with nanometre precision<br />
to be probed [54]. This nanothermal cantilever has a conductive coating<br />
through which an electrical current is passed to an integrated heater<br />
located directly above the probe. By varying the resistance of the circuit,<br />
the temperature of the heater can be controlled up to 500°C, depending<br />
on the choice of probe. When the probe is in contact with the surface, any<br />
deflections in the cantilever are recorded <strong>and</strong> this can reveal the nature<br />
of the material (e.g. amorphous versus crystalline) [55, 56] <strong>and</strong> the occurrence<br />
of thermal phase transitions such as melting or glass transitions [54].<br />
The development of the nanoprobe has also given the ability to maintain<br />
a constant probe temperature during scanning to allow for thermal imaging<br />
of the surface or for controlled thermal nanolithography [57, 58]. An<br />
example of such local thermal analysis is shown in Figure 6.9, where an<br />
Deflection (arb. units)<br />
Dehydration<br />
Melt<br />
50 100 150 200<br />
Temperature (°C)<br />
FIgure 6.9 Nanoscale thermal analysis (NTA) of lactose monohydrate. (Left) Image<br />
recorded with an NTA probe after local thermal analysis has been performed at the area, highlighted<br />
with an arrow. The nanoscale pit is the result of local melting. (Right) Corresponding<br />
cantilever deflection trace, revealing dehydration <strong>and</strong> melting of the lactose monohydrate.
190 6. NANOSCALE ANALySIS Of PHARMACEUTICALS by SCANNINg PRObE MICROSCOPy<br />
individual lactose monohydrate crystal within a blend has been studied.<br />
The resultant indent due to local melting is apparent in the image (now<br />
with nanometre resolution due to the sharper probe). The corresponding<br />
cantilever deflection trace reveals the loss of water from the monohydrate<br />
<strong>and</strong> the melt of the anhydrous form at temperatures consistent with previous<br />
SThM <strong>and</strong> bulk analysis [59].<br />
6.5 CoNCluSIoNS<br />
As the barriers to developing new medicines become ever greater<br />
owing to new challenges in delivery, greater regulation <strong>and</strong> the scarcity<br />
of ‘easy’ molecules to develop into medicines, the role of probe microscopes<br />
in pharmaceuticals development is set to increase. The advantages<br />
of nanoscale resolution, minimal sample preparation <strong>and</strong> the non-invasive<br />
imaging capabilities of AFM in particular make this an ideal tool to bring<br />
new approaches to the investigation of drug material properties.<br />
This chapter has provided an insight into the improved underst<strong>and</strong>ing<br />
of pharmaceutical drug development available by AFM <strong>and</strong> thermal<br />
versions of AFM. These approaches have in a relatively short period of<br />
time provided a series of new insights into drug particle interactions<br />
<strong>and</strong> formulation structure. The capacity of AFM to not only image at the<br />
nanoscale but also to investigate interactions <strong>and</strong> to spatially map surface<br />
properties <strong>and</strong> to operate in a variety of environments consistent<br />
with pharmaceutical testing, manufacture <strong>and</strong> delivery provides great<br />
opportunity.<br />
ACkNowledgeMeNTS<br />
I would like to thank all who have contributed to this research, especially<br />
Martyn Davies, Jennifer Hooton, Ardeshir Danesh, Jin Zheng, Jeff<br />
James, Matthew Bunker, Michael Davies <strong>and</strong> Anne Turner.<br />
ABBrevIATIoNS ANd SyMBolS<br />
DPI dry powder inhaler<br />
Es elastic modulus of sample N m�2 K combined elastic modulus of tip <strong>and</strong> sample N m�2 L load force N
LTA local thermal analysis<br />
PTFE polytetrafluoroethylene<br />
R radius of probe m<br />
SThM scanning thermal microscopy<br />
T g glass transition temperature °C<br />
v s<br />
v t<br />
Poisson’s ratio for sample<br />
Poisson’s ratio of tip<br />
� indentation depth m<br />
references<br />
AbbREvIATIONS ANd SyMbOLS 191<br />
[1] I. Verweire, E. Schacht, B.P. Qiang, K. Wang, I. De Scheerde, Evaluation of fluorinated<br />
polymers as coronary stent coating, J. Mater. Sci. Mater. Med. 11 (2000) 207–212.<br />
[2] A.L. Lewis, L.A. Tolhurst, P.W. Stratford, Phosphorylcholine-coated stent, J. Long<br />
Term Eff. Med. Implants 12 (2002) 231–250.<br />
[3] D. Traini, P. Yong, P. Rogueda, R. Price, The use of AFM <strong>and</strong> surface energy measurements<br />
to investigate drug-canister material interactions in a model pressurized<br />
metered dose inhaler formulation, Aerosol Sci. Technol. 40 (2006) 227–236.<br />
[4] A. Danesh, M.C. Davies, S. Hinder, C.J. Roberts, S.J.B. Tendler, P.M. Williams,<br />
M.J. Wilkins, Surface characterization of aspirin crystal planes by dynamic chemical<br />
force microscopy, Anal. Chem. 72 (2000) 3419–3422.<br />
[5] A. Danesh, S.D. Connell, M.C. Davies, S.J.B. Tendler, C.J. Roberts, P.M. Williams,<br />
M.J. Wilkins, An in situ dissolution study of aspirin crystal planes (100) <strong>and</strong> (001) by<br />
atomic force microscopy, Pharm. Res. 18 (2001) 299–303.<br />
[6] M. Davies, A. Brindley, X. Chen, M. Marlow, S.W. Doughty, I. Shrubb, C.J. Roberts,<br />
Characterization of drug particle surface energetics <strong>and</strong> Young’s modulus by atomic<br />
force microscopy <strong>and</strong> inverse gas chromatography, Pharm. Res. 22 (2005) 1158–1166.<br />
[7] R.I. Larsen, The adhesion <strong>and</strong> removal of particles attached to air filter surfaces, Am.<br />
Indust. Hyg. J. 19 (1958) 265–270.<br />
[8] F. Podczeck, The development of a cascade impactor simulator based on adhesion<br />
force measurements to aid the development of dry powder inhalations, Chem. Pharm.<br />
Bull. 45 (1997) 911–917.<br />
[9] B.C. Hancock, S.D. Clas, K. Christensen, Micro-scale measurement of the mechanical<br />
properties of compressed pharmaceutical powders. 1: The elasticity <strong>and</strong> fracture behavior<br />
of microcrystalline cellulose, I. J. Pharm. 209 (2000) 27–35.<br />
[10] S. Bin Baie, J.M. Newton, F. Podczeck, The characterization of the mechanical properties<br />
of pharmaceutical materials, Eur. J. Pharm. Biopharm. 42 (1996) 138–141.<br />
[11] G. Binning, C.F. Quate, Ch. Geber, Atomic force microscope, Phys. Rev. Lett. 56 (1986)<br />
930–933.<br />
[12] N. Jalili, K. Laxminarayana, A review of atomic force microscopy imaging systems:<br />
application to molecular metrology <strong>and</strong> biological sciences, Mechatronics 14 (2004)<br />
907–945.<br />
[13] N.H. Green, S. Allen, M.C. Davies, C.J. Roberts, S.J.B. Tendler, P.M. Williams, Force<br />
sensing <strong>and</strong> mapping by atomic force microscopy, Trends Analyt. Chem. 21 (2002)<br />
64–73.<br />
[14] W.A. Ducker, T.J. Senden, R.M. Pashley, Direct measurement of colloidal forces using<br />
an atomic force microscope, Nature 353 (1991) 239–241.
192 6. NANOSCALE ANALySIS Of PHARMACEUTICALS by SCANNINg PRObE MICROSCOPy<br />
[15] T.H. Ibrahim, T.R. Burk, F.M. Etzler, R.D. Neuman, Direct adhesion measurements of<br />
pharmaceutical particles to gelatin capsule surfaces, J. Adhes. Sci. Technol. 14 (2000)<br />
1225–1242.<br />
[16] C.T. Gibson, D.A. Smith, C.J. Roberts, Calibration of silicon atomic force microscope<br />
cantilevers, Nanotechnology 16 (2005) 234–238.<br />
[17] M.D. Louey, P. Mulvaney, P.J. Stewart, Characterisation of adhesional properties<br />
of lactose carriers using atomic force microscopy, J. Pharm. Biomed. Anal. 25 (2001)<br />
559–567.<br />
[18] J.K. Eve, N. Patel, S.Y. Luk, S.J. Ebbens, C.J. Roberts, A study of single drug particle<br />
adhesion interactions using atomic force microscopy, Int. J. Pharm. 238 (2002) 17–27.<br />
[19] V. Berard, E. Lesniewska, C. Andres, D. Pertuy, C. Laroche, Y. Pourcelot, Affinity scale<br />
between a carrier <strong>and</strong> a drug in DPI studied by atomic force microscopy, I. J. Pharm.<br />
247 (2002) 127–137.<br />
[20] P.M. Young, R. Price, M.J. Tobyn, M. Buttrum, F. Dey, The influence of relative humidity<br />
on the cohesion properties of micronized drugs used in inhalation therapy, J. Pharm.<br />
Sci. 93 (2004) 753–761.<br />
[21] J.C. Hooton, C.S. German, S. Allen, M.C. Davies, C.J. Roberts, S.J.B. Tendler,<br />
P.M. Williams, An atomic force microscopy study of the effect of nanoscale contact<br />
geometry <strong>and</strong> surface chemistry on the adhesion of pharmaceutical particles, Pharm.<br />
Res. 21 (2004) 953–961.<br />
[22] J.C. Hooton, C.S. German, S. Allen, M.C. Davies, C.J. Roberts, S.J.B. Tendler,<br />
P.M. Williams, Characterisation of particle-interactions by atomic force microscopy:<br />
the effect of contact area, Pharm. Res. 20 (2003) 508–514.<br />
[23] R. Ashayer, P.F. Luckham, S. Manimaaran, P. Rogueda, Investigation of the molecular<br />
interactions in a pMDI formulation by atomic force microscopy, Eur. J. Pharm. Sci. 21<br />
(2004) 533–543.<br />
[24] P.M. Young, R. Price, D. Lewis, S. Edge, D. Traini, Under pressure: predicting pressurized<br />
metered dose inhaler interactions using the atomic force microscope, J. Colloid<br />
Interface Sci. 262 (2003) 298–302.<br />
[25] M. Pierce, J. Stuart, A. Pungor, P. Dryden, V. Hlady, Adhesion force measurements<br />
using an atomic-force microscope upgraded with a linear position-sensitive detector,<br />
Langmuir 10 (1994) 3217–3221.<br />
[26] P. Begat, D.A. Morton, R. Price, J.N. Staniforth, Investigation into the cohesive–<br />
adhesive force balance within a dry powder inhaler formulation, Res. Drug Deliv. IX<br />
3 (2003) 729–732.<br />
[27] M.J. Bunker, C.J. Roberts, M.C. Davies, M.B. James, A nanoscale study of particle friction<br />
in a pharmaceutical system, Int. J. Pharm. 325 (2006) 163–171.<br />
[28] M.J. Bunker, M.C. Davies, X. Chen, M.B. James, C.J. Roberts, Single particle friction on<br />
blister packaging materials used in dry powder inhalers, Eur. J. Pharm. Sci. 29 (2006)<br />
405–413.<br />
[29] B.C. Hancock, G.T. Carlson, D.D. Ladipo, B.A. Langdon, M.P. Mullarney, Comparison<br />
of the mechanical properties of the crystalline amorphous forms of a drug substance,<br />
Int. J. Pharm. 241 (2002) 73–85.<br />
[30] X. Liao, T.S. Wiedmann, Measurement of process-dependent material properties of<br />
pharmaceutical solids by nanoindentation, J. Pharm. Sci. 94 (2004) 79–92.<br />
[31] C.C. Kwana, Y.Q. Chena, Y.L. Dinga, D.G. Papadopoulosb, A.C. Benthamb,<br />
M. Ghadiria, Development of a novel approach towards predicting the milling behaviour<br />
of pharmaceutical powders, Eur. J. Pharm. Sci. 23 (2004) 327–336.<br />
[32] S. Jain, Mechanical properties of powders for compaction <strong>and</strong> tableting: an overview,<br />
Pharm. Sci. Technol. Today 2 (1999) 20–31.<br />
[33] P. Narayan, B.C. Hancock, The relationship between the particle properties, mechanical<br />
behavior, <strong>and</strong> surface roughness of some pharmaceutical excipient compacts, Mater.<br />
Sci. Eng. A 355 (2003) 24–36.
REfERENCES 193<br />
[34] V.M. Mastersona, X. Cao, Evaluating particle hardness of pharmaceutical solids using<br />
AFM nanoindentation, Pharm. Nanotechnol. 362 (2008) 163–171.<br />
[35] X. Liao, T.S. Wiedmann, Characterization of pharmaceutical solids by scanning probe<br />
microscopy, J. Pharm. Sci. 93 (2004) 2250–2258.<br />
[36] W.C. Oliver, G.M. Pharr, An improved technique for determining hardness <strong>and</strong> elastic<br />
modulus using load <strong>and</strong> displacement sensing indentation experiments, J. Mater. Res.<br />
7 (1992) 1564–1583.<br />
[37] S. Ward, M. Perkins, J. Zhang, C.J. Roberts, C.E. Madden, S.Y. Luk, N. Patel,<br />
S.J. Ebbens, Identifying <strong>and</strong> mapping surface amorphous domains, Pharm. Res. 22<br />
(2005) 1195–1202.<br />
[38] M. Radmacher, Measuring the elastic properties of biological samples with the AFM,<br />
IEEE Eng. Med. Biol. Mag. 16 (1997) 47–57.<br />
[39] R.W. Carpick, D.F. Ogletree, M. Salmeron, A general equation for fitting contact area<br />
<strong>and</strong> friction vs load measurements, J. Colloid Interface Sci. 211 (1999) 395–400.<br />
[40] J.J. Wang, T.L. Li, S.D. Bateman, R. Erck, K.R. Morris, Modeling of adhesion in tablet<br />
compression-I. Atomic force microscopy <strong>and</strong> molecular simulation, J. Pharm. Sci. 92<br />
(2003) 798–814.<br />
[41] A.V. Nguyen, J. Nalaskowski, J.D. Miller, A study of bubble–particle interaction using<br />
atomic force microscopy, Miner. Eng. 16 (2003) 1173–1181.<br />
[42] J.K. Hobbs, C. Vasilev, A.D.L. Humphries, Real time observation of crystallization<br />
in polyethylene oxide with video rate atomic force microscopy, Polymer 46 (2005)<br />
10226–10236.<br />
[43] R. Price, P.M. Young, Visualization of the crystallisation of lactose from the amorphous<br />
state, J. Pharm. Sci. 93 (2004) 155–164.<br />
[44] T. Li, K.R. Morris, K. Park, Influence of solvent <strong>and</strong> crystalline supramolecular<br />
structure on the formation of etching patterns on acetaminophen single crystals: a<br />
study with atomic force microscopy <strong>and</strong> computer simulation, J. Phys. Chem. B 104<br />
(2000) 2019–2032.<br />
[45] K.M. Shakesheff, M.C. Davies, A. Domb, D.E. Jackson, C.J. Roberts, S.J.B. Tendler,<br />
P.M. Williams, In situ atomic force microscopy visualization of the degradation of<br />
melt-crystallized poly(sebacic anhydride), Macromolecules 8 (1995) 1108–1114.<br />
[46] C. Thompson, M.C. Davies, C.J. Roberts, S.J.B. Tendler, M.J. Wilkinson, The effects of<br />
additives on the growth <strong>and</strong> morphology of paracetamol (acetaminophen) crystals,<br />
Int. J. Pharm. 280 (2004) 137–150.<br />
[47] X. Chen, M.C. Davies, C.J. Roberts, S.J.B. Tendler, P.M.W. Williams, N.A. Burnham,<br />
Optimizing phase imaging via dynamic force curves, Surf. Sci. 460 (2000) 292–300.<br />
[48] A. Danesh, X. Chen, M.C. Davies, C.J. Roberts, G.H.W. S<strong>and</strong>ers, S.J.B. Tendler,<br />
P.M. Williams, M.J. Wilkins, The discrimination of drug polymorphic forms from<br />
single crystals using atomic force microscopy, Pharm. Res. 17 (2000) 887–890.<br />
[49] E. Shen, R. Pizsczek, B. Dziadul, B. Narasimhan, Microphase separation in bioerodible<br />
copolymers for drug delivery, Biomaterials 22 (2001) 201–210.<br />
[50] D.M. Price, M. Reading, A. Hammiche, H.M. Pollock, Micro-thermal analysis: scanning<br />
thermal microscopy <strong>and</strong> localised thermal analysis, Int. J. Pharm. 192 (1999) 85–96.<br />
[51] K. Six, J. Murphy, I. Weuts, D.Q.M. Craig, G. Verreck, J. Peeters, M. Brewster, G. Van<br />
den Mooter, Identification of phase separation in solid dispersions of Itraconazole <strong>and</strong><br />
Eudragit® E100 using microthermal analysis, Pharm. Res. 20 (2003) 135–138.<br />
[52] G.H.W. S<strong>and</strong>ers, C.J. Roberts, A. Danesh, D. Craig, A. Murray, M.C. Davies, S.J.B. Tendler,<br />
P.M. Williams, Discrimination of polymorphic forms of a drug product by localized thermal<br />
analysis, J. Microsc. 198 (2000) 77–81.<br />
[53] A. Bauer-Br<strong>and</strong>l, Polymorphic transition of cimetidine during manufacture of solid<br />
dosage forms, Int. J. Pharm. 140 (1996) 195–206.<br />
[54] B.A. Nelson, W.P. King, Measuring material softening with nanoscale spatial resolution<br />
using heated silicon probes, Rev. Sci. Instrum. 78 (2007) 023702.
194 6. NANOSCALE ANALySIS Of PHARMACEUTICALS by SCANNINg PRObE MICROSCOPy<br />
[55] P.G. Royall, V.L. Kett, C.S. Andrews, D.Q.M. Craig, Identification of crystalline <strong>and</strong><br />
amorphous regions in low molecular weight materials using microthermal analysis,<br />
J. Phys. Chem. B 105 (2001) 7021–7026.<br />
[56] L. Harding, W.P. King, X. Dail, D.Q.M. Craig, M. Reading, Nanoscale characterisation<br />
<strong>and</strong> imaging of partially amorphous materials using local thermomechanical analysis<br />
<strong>and</strong> heated tip AFM, Pharm. Res. 24 (2007) 2048–2054.<br />
[57] T-H. Fang, W-J. Chang, Microthermal machining using scanning thermal microscopy,<br />
Appl. Surf. Sci. 240 (2005) 312–317.<br />
[58] W.P. King, S. Saxena, B.A. Nelson, B.L. Weeks, R. Pitchimani, Nanoscale thermal analysis<br />
of an energetic material, Nano Lett. 6 (2006) 2145–2149.<br />
[59] L. Bond, S. Allen, M.C. Davies, C.J. Roberts, A.P. Shivji, S.J.B. Tendler, P.M. Williams,<br />
J. Zhang, Differential scanning calorimetry <strong>and</strong> scanning thermal microscopy analysis<br />
of pharmaceutical materials, Int. J. Pharm. 243 (2002) 71–82.
C H A P T E R<br />
7<br />
Micro/Nanoengineering <strong>and</strong><br />
AFM for Cellular Sensing<br />
Huabing Yin, Gordon McPhee <strong>and</strong><br />
Phil S. Dobson<br />
O U T L I N E<br />
7.1 Introduction 195<br />
7.1.1 How Do Cells Respond to the ECM? 196<br />
7.2 Engineering the ECM for Probing Cell Sensing 198<br />
7.2.1 Surface Patterning (Chemical Signals) 199<br />
7.2.2 Nanotopography 205<br />
7.2.3 Nanoscale Measurement: Challenges <strong>and</strong> Opportunities for AFM 209<br />
7.3 AFM in Cell Measurement 211<br />
7.3.1 AFM Imaging of Cells 211<br />
7.3.2 Elasticity Measurement of Living Cells 214<br />
7.4 Conclusions 217<br />
Acknowledgements 219<br />
Abbreviations <strong>and</strong> Symbols 219<br />
References 220<br />
7.1 IntroduCtIon<br />
Cells live in a complex extracellular matrix (ECM), which consists of<br />
neighbouring cells, proteins <strong>and</strong> extracellular fluids, inside the bodies of<br />
animals <strong>and</strong> plants. Cells constantly monitor the chemical <strong>and</strong> physical<br />
signals from their surroundings <strong>and</strong> react accordingly. A large body of<br />
Atomic Force Microscopy in Process Engineering 195<br />
© 2009, Elsevier Ltd
196 7. MICRO/NANOENgINEERINg ANd AFM FOR CELLULAR SENSINg<br />
evidence has shown that the various interactions between a cell <strong>and</strong> different<br />
microenvironments play crucial roles in embryonic morphogenesis,<br />
tissue formation <strong>and</strong> maintenance of physiological functions [1]. Perturbations<br />
of cellular microenvironments or the adhesion of cells to the ECM<br />
can cause genetic defects, autoimmune diseases <strong>and</strong> cancers [2, 3]. It is<br />
also well known that in cancer metastasis, malignant cancer cells are able<br />
to break down tissue architecture <strong>and</strong> invade distant organ sites [4, 5].<br />
The ECM is an intricate network within which biomolecules are<br />
precisely organised [6]. Classes of structural ECM proteins, mainly, collagen,<br />
glycoproteins <strong>and</strong> proteoglycans, form highly organised nanoscale<br />
structures, providing cells with both biological information <strong>and</strong> physical<br />
scaffolds for adhesion <strong>and</strong> migration. Although the regulatory functions<br />
of soluble factors (i.e. growth factors <strong>and</strong> hormones) present in the ECM<br />
have been well investigated, it has recently become increasingly recognised<br />
that the physics <strong>and</strong> mechanics of the ECM also have a significant<br />
impact on cell function <strong>and</strong> fate. Revealing the precise mechanisms underlying<br />
these processes is intrinsically challenging – the events happen at<br />
many dimensions from single molecules to the macrotissue level, <strong>and</strong> in a<br />
dynamic/collective <strong>and</strong> hierarchical manner.<br />
In order to better underst<strong>and</strong> the interactions between the cells <strong>and</strong><br />
the ECM <strong>and</strong> subsequent responses, engineered substrates <strong>and</strong> scaffolds<br />
are being developed to replace the native ECM. This has required inputs<br />
from many fields, ranging from surface engineering, material science <strong>and</strong><br />
tissue engineering to cell biology. In this chapter, we will describe current<br />
developments of micro- <strong>and</strong> nanoengineered ECM materials <strong>and</strong><br />
structures for the investigation of adhesion-associated responses of cells<br />
to chemical <strong>and</strong> mechanical cues. These efforts will be illustrated by an<br />
interconnected set of examples.<br />
Importantly, as the length scales being examined in these studies have<br />
progressively shrunk, progress in interpreting the nanoworld has become<br />
increasingly dependent on techniques capable of nanoscale measurements<br />
within physiologically relevant environments. In this context, atomic<br />
force microscopy (AFM) has emerged as a powerful <strong>and</strong> multifunctional<br />
nanoscale tool, opening exciting new possibilities to address mechanistic<br />
questions in cell biology that may facilitate the development of efficient<br />
therapies for human health. In the following sections, we briefly introduce<br />
the basic biological principles for adhesion-associated cell sensing,<br />
followed by the engineering methods involved in generating substrates<br />
<strong>and</strong> materials to study cellular interactions, <strong>and</strong> finally the methodology<br />
associated with AFM measurements on these systems.<br />
7.1.1 How do Cells respond to the ECM?<br />
First, we review two important subcellular systems that are of fundamental<br />
importance in cell adhesion to the ECM: the cell membrane <strong>and</strong>
7.1 INTROdUCTION 197<br />
the cytoskeleton (Figure 7.1) (for background reading see, Alberts et al. [7]).<br />
The cell membrane is a barrier that separates the interior of the cell from<br />
the outside environment; it regulates the transport of molecules into <strong>and</strong><br />
out of the cell <strong>and</strong> maintains the interior of the cell at optimal levels of pH<br />
<strong>and</strong> ionic concentrations. Cell membranes are primarily made of a selectively<br />
permeable lipid bilayer, containing various functional proteins that<br />
are involved in a range of specific cellular activities. For example, 25–50%<br />
of membrane receptors may be adhesive receptors [8]. The interior compartment<br />
next to the cell membrane is the cytoplasm. This accommodates<br />
a number of specialised subcellular organelles that cooperate to maintain<br />
cell function. The cytoskeleton, which is located within the cytoplasm, is<br />
made up of three types of long rod-shaped molecules: microfilaments (e.g.<br />
actin stress fibre), microtubules (e.g. tubulin) <strong>and</strong> intermediate filaments<br />
(e.g. vimentin). These molecules attach to one another, link to other subcellular<br />
systems, such as the cell membrane <strong>and</strong> cell nucleus, <strong>and</strong> build<br />
a framework to give the cell both shape <strong>and</strong> movement. The configuration<br />
of the cytoskeleton dynamically adapts during cellular processes <strong>and</strong><br />
undergoes microscopically observable morphological changes.<br />
It is now clear that a particular family of transmembrane cell surface<br />
receptors, the integrins, mediate many of the interactions between a cell<br />
<strong>and</strong> the ECM. They both recognise peptide sequences, such as Arg-Gly-Asp<br />
(RGD) within the chains of certain ECM proteins (e.g. fibronectin), <strong>and</strong><br />
connect the cytoskeleton to the ECM. During this process, adhesive<br />
contacts between the cell <strong>and</strong> the ECM are formed [9]. A common type<br />
of adhesive contact involves multiprotein complexes, called focal adhesions.<br />
These comprise integrins, the associated cytoplasmic proteins,<br />
<strong>and</strong> a number of protein kinases [1, 10]. Focal adhesions are the major<br />
sites for actin stress fibre attachment <strong>and</strong> thus a connection between the<br />
cytoskeleton <strong>and</strong> the ECM. Integrins that are bound to the ECM transmit<br />
Nucleus<br />
Focal adhesion<br />
Cell membrane<br />
F-actin stress fibre<br />
Direction of motion<br />
ECM or substrate<br />
Microtubule<br />
Filopodium<br />
Pseudopodium<br />
Lamellopodium<br />
FIgurE 7.1 Schematic drawing of cell adhesion to an ECM or substrate. The cell<br />
adheres firmly to the ECM through focal adhesions (a multiprotein complex). The focal<br />
adhesions are the sites for the attachment of F-actin stress fibres – one type of cytoskeleton<br />
protein. Filopodium <strong>and</strong> lamellopodium are located at the leading edge for cell to migrate.
198 7. MICRO/NANOENgINEERINg ANd AFM FOR CELLULAR SENSINg<br />
mechanical stress across the plasma membrane <strong>and</strong> convey the traction<br />
force that develops in the cytoskeleton to the ECM. Unbound integrins<br />
are mobile within the cell membrane <strong>and</strong> readily form clusters <strong>and</strong> focal<br />
adhesions in a tension-dependent manner [11]. Integrins also participate<br />
in other signalling transductions that regulate cell growth [12].<br />
During the initial phase of cell adhesion, interactions between the<br />
integrins <strong>and</strong> the ECM lig<strong>and</strong>s are independent of force, but rapidly the<br />
resultant adhesion induces the activation of Rac <strong>and</strong> Cdc42 protein pathways,<br />
leading to the formation of filopodia <strong>and</strong> lamellipodia (Figure 7.1).<br />
These structures create a small adhesion site, called a focal complex (in<br />
the order of ~1 �m). From this point, cells start to exert traction forces on<br />
the ECM, with about 0.8–0.9 nN �m �2 being exerted by lamellipodia [10].<br />
Filopodia are effectively the “antennae” of the cell, formed at the leading<br />
edge. Focal adhesions <strong>and</strong> focal complexes recruit the same core proteins<br />
[13]; however, focal adhesions are much larger in size <strong>and</strong> integrins packing<br />
density, <strong>and</strong> produce larger forces of a few nanonewton per square<br />
micrometre [14]. Focal complexes can mature into focal adhesions if there<br />
is an increase in force at the adhesion site [10, 15, 16], or by the activation<br />
of the Rho pathway [13]. Thus, physical tension between a cell <strong>and</strong> the<br />
ECM appears to be essential for focal adhesion formation <strong>and</strong> any subsequent<br />
firm adhesion.<br />
Finally, it should be noted that cell adhesion, spreading <strong>and</strong> migration<br />
require assembly <strong>and</strong> disassembly of multiple focal adhesions. This is<br />
regulated by integrin–lig<strong>and</strong> binding events <strong>and</strong> can be stimulated by<br />
properties associated with the ECM (e.g. lig<strong>and</strong> densities <strong>and</strong> stiffness<br />
of the matrix) as well as by intracellular signals [9, 17, 18]. To date, the<br />
detailed mechanisms that regulate the organisation of these adhesive<br />
complexes are yet to be elucidated. However, abundant evidence suggests<br />
a highly dynamic feedback loop between a cell <strong>and</strong> its microenvironment,<br />
which is constantly modulated by delicate changes in a vast range of<br />
(bio)chemical <strong>and</strong> physical parameters.<br />
7.2 EngInEErIng tHE ECM For ProbIng<br />
CEll SEnSIng<br />
A typical animal cell is �10–100 �m in size. The major components of<br />
focal adhesion sites, such as integrins, talin <strong>and</strong> vinculin, are proteins with<br />
dimensions in the range of several nanometres. Both of these size scales<br />
require sophisticated tools for visualisation <strong>and</strong> manipulation of these<br />
functional components. However, the local microenvironments of cells are<br />
inherently heterogeneous <strong>and</strong> dynamically remodelled at different stages<br />
in the life cycle of a cell. All of these factors can lead to great uncertainty<br />
<strong>and</strong> variability in the interpretation of observations. Nevertheless, it is
7.2 ENgINEERINg THE ECM FOR PRObINg CELL SENSINg 199<br />
because of these challenges that intense effort from many fields has been<br />
attracted, leading to the combination of micro- <strong>and</strong> nanotechnology with<br />
biological sciences <strong>and</strong> the formation of the interdisciplinary ‘bionanotechnology’<br />
field.<br />
Advances in micro- <strong>and</strong> nanofabrication can produce precisely controlled<br />
model systems at a single molecule level, <strong>and</strong> provide a systematic<br />
approach to dissect the roles of intertwined parameters in the ECM. In<br />
combination with other approaches (including optical microscopy, AFM<br />
<strong>and</strong> scanning electron microscopy [SEM]), these fabrication methods have<br />
proven to be powerful in the elucidation of complex cell behaviour. Here<br />
we will discuss the development of engineered substrates for the investigation<br />
of cell interactions with the ECM. Specifically, we will review the<br />
achievements of these substrates in mimicking chemical <strong>and</strong> physical cues<br />
in the ECM. Although not explicitly stated in the review below, many of<br />
these studies described have been underpinned by AFM measurements of<br />
surface interactions, topography or materials compliance.<br />
7.2.1 Surface Patterning (Chemical Signals)<br />
Micropatterning<br />
Micropatterning is based on photolithography [19], which produces<br />
features with dimensions over 1 �m. As illustrated in the schematic representation<br />
in Figure 7.2(A), this process involves the UV irradiation of a<br />
spin-coated photosensitive polymer layer (photoresist) through a mask.<br />
UV irradiation through the mask causes the photoresist polymer chains<br />
to either break up (positive resist) or crosslink (negative resist), leading<br />
to a difference in solubility between the exposed <strong>and</strong> unexposed regions<br />
when immersed in a “developer” solution. After developing, the patterns<br />
from the mask have been effectively transferred onto the substrate. The<br />
patterned photoresist can then serve as a protecting layer in subsequent<br />
processes, such as lift-off to produce metal patterns, or etching to generate<br />
a relief on the substrate.<br />
Micropatterning has been extensively used for surface patterning of<br />
biological molecules. Its main purpose is to allow fine control over the size<br />
<strong>and</strong> spatial arrangement of regions that can be specifically functionalised<br />
for the attachment of ECM proteins. For this, selective immobilisation of<br />
adhesive molecules on the patterned area is required. It should be noted<br />
that preventing the physisorption of biomolecules (especially proteins)<br />
on both the patterned <strong>and</strong> non-patterned surfaces is equally important, as<br />
non-specifically adsorbed proteins could also serve as an adhesive region<br />
for cell attachment.<br />
A key development in the generation of chemically patterned substrates<br />
has exploited the formation of self-assembly monolayers (SAM) from<br />
heterobifunctional organic molecules that bear specific functional groups
200 7. MICRO/NANOENgINEERINg ANd AFM FOR CELLULAR SENSINg<br />
(A)<br />
(B)<br />
Photolithography<br />
(a)<br />
(b)<br />
(c)<br />
(d)<br />
Mask<br />
Etching Lift-off<br />
(e) (g)<br />
(f)<br />
(h)<br />
Photoresist<br />
Substrate<br />
(a) (b) (c)<br />
(f) (e) (d)<br />
Metal<br />
FIgurE 7.2 Schematic diagram of methods used in micropatterning. (A) Photolithography<br />
<strong>and</strong> (B) microcontact printing. In photolithography a mask with opaque <strong>and</strong> transparent<br />
features is brought into contact with a substrate coated with a photosensitive polymer (photoresist)<br />
(a <strong>and</strong> b). UV light is shone through the mask, exposing the polymer beneath the<br />
transparent regions of the mask (c). The mask is removed <strong>and</strong> the resist developed, removing<br />
either the exposed or unexposed regions of the resist depending on the resist type (d).<br />
The resist layer can be used as a protective mask during a process to etch the unprotected<br />
regions of the substrate (e)–(f). Alternatively, metal can be evaporated onto the sample,<br />
remaining on the regions of exposed substrate after the resist has been dissolved in a ‘lift-off’<br />
process (g)–(h). In microcontact printing, a topographically patterned stamp is ‘inked’ by<br />
contacting it with a sample coated with the desired chemical species (a)–(c). When the inked<br />
stamp is brought into contact with a clean substrate, the species are deposited on the surface<br />
in patterned regions dictated by the topography of the stamp (d)–(f).
7.2 ENgINEERINg THE ECM FOR PRObINg CELL SENSINg 201<br />
which react with the substrate. Upon SAM formation, the chemical properties<br />
of the surface are no longer determined by the underlying substrate<br />
but by the exposed functional tail groups at the non-substrate end of the<br />
SAM. A commonly used system is the self-assembled monolayer of an<br />
alkanethiol (SH(CH 2) nX) on gold; the sulfhydryl head groups (SH–) spontaneously<br />
form sulphur–gold bonds <strong>and</strong> leave the functional tail group<br />
structures (X) tethered away from the surface. The terminal X groups can<br />
be tailored to have different functionalities [20], e.g., NH 2– <strong>and</strong> COOH–<br />
groups for subsequent attachment of proteins or peptides via common<br />
bioconjugation chemistries [21]. The use of mixed or diluted monolayers<br />
<strong>and</strong> tuning the length of the hydrocarbon spacer chain can also be used to<br />
achieve desired surface properties [22]. SAMs of alkylsilane on hydroxylated<br />
surfaces of SiO 2/Si substrates, such as glass <strong>and</strong> silica, have also been<br />
extensively developed, although here, greater care needs to be taken over<br />
the reaction conditions to avoid self-polymerisation on the surface.<br />
As photolithography requires clean room facilities to prevent dust particles<br />
spoiling the pattern transfer process, the infrastructure costs mean that<br />
these techniques are often not easily accessible on a regular basis. Thus, a<br />
pioneering method has been developed by Whitesides’ group, called ‘soft<br />
lithography’ (microcontact printing, Figure 7.2(B)). This uses a physical<br />
stamp made from poly(dimethysiloxne) (PDMS) elastomer [23]. The stamp<br />
is a negative replica of a microstructured substrate (master) made using<br />
conventional photolithography <strong>and</strong>/or etching methods. Generally, the<br />
stamp is fabricated by thermal curing of PDMS against a master template<br />
followed by peeling away the cured structure. The PDMS stamps can then<br />
be inked with silanes, alkanethiols or ECM proteins to produce patterns<br />
on a wide range of substrates [24]. Unstamped regions can be blocked to<br />
resist non-specific protein adsorption by various methods such as the use<br />
of PEG-terminated SAMs or rinsing with denatured albumin solution. The<br />
PDMS stamps can be repeatedly replicated from the master <strong>and</strong> reused<br />
many times, thereby significantly reducing the cost of fabrication. Using<br />
this methodology, microcontact printing has greatly extended the range of<br />
substrates that can be engineered for use in cell studies since many of the<br />
polymeric materials compatible with biological studies are not suitable for<br />
patterning using traditional photolithography processes.<br />
An early example of the use of micropatterned surfaces is one which<br />
allowed systematic examination of the interaction of cells with substrate<br />
surface chemistries <strong>and</strong> ECM components [25]. Observations of cell<br />
growth on metallic <strong>and</strong> polymeric micropatterns have provided invaluable<br />
information to develop new types of cell culture vessels <strong>and</strong> screen<br />
the biocompatiblity of metal materials for implants. Cell function data on<br />
SAM patterns have enabled the identification of chemistries <strong>and</strong> systems<br />
that possess designed properties; e.g. observations from patterns made<br />
using a range of poly(ethylene glycol) (PEG)-derivatised alkanethiols
202 7. MICRO/NANOENgINEERINg ANd AFM FOR CELLULAR SENSINg<br />
(<strong>and</strong> silanes) have found which exhibit the lowest protein physisorption<br />
(<strong>and</strong> hence block cell adhesion) [26–28]. These SAMs can therefore be<br />
used to form a non-adhesive background. By the integration of specific<br />
adhesion lig<strong>and</strong>s into these patterns, defined spatial distribution of the<br />
lig<strong>and</strong>s can be formed, as shown in Figure 7.3(a). Here, a pattern of fluorescently<br />
labelled fibronectin on a PEG-passivated surface was prepared<br />
by photolithography. Figure 7.3(b) demonstrates that fibroblasts specifically<br />
attach <strong>and</strong> proliferate in the adhesive isl<strong>and</strong>s.<br />
Cellular interactions in vivo are largely varied. Different types of cells<br />
are juxtaposed in the ECM <strong>and</strong> each secretes different cues at different<br />
times. Thus, cells are exposed to temporally spatially regulated concentrations<br />
or dynamically changing gradients of adhesive cues (e.g. due<br />
to diffusion of soluble molecules emanating from particular cells). With<br />
surface engineering being able to tune spatial distances <strong>and</strong> pattern sizes,<br />
micropatterning provides a convenient way to study how cells sense<br />
the spatial distribution of adhesive cues. A study of cells on a micropattern<br />
of ECM lig<strong>and</strong>s has discovered that the amount of integrin does not<br />
always determine cell life or death, but instead this is often indicated by<br />
the degree of cell spreading <strong>and</strong> its shape [29]. Surface patterning of different<br />
lig<strong>and</strong>s for two types of cells has also advanced the in vitro study<br />
100 μm<br />
(a)<br />
200 μm<br />
(b)<br />
FIgurE 7.3 Example of surface patterning for the generation of adhesive protein patterns.<br />
(a) Fluorescence-labelled fibronectin patterns on a glass substrate generated using<br />
photolithography. (b) 3T3 fibroblast specifically attached <strong>and</strong> growing on the collagen patterns<br />
but not on the PEG surface between the patterns.
7.2 ENgINEERINg THE ECM FOR PRObINg CELL SENSINg 203<br />
of parenchymal cells such as hepatocytes (liver cells) which require heterotypic<br />
interactions between non-parenchymal cells (e.g. fibroblast) to<br />
maintain their liver cell phenotype [30]. In this study, the microfabrication<br />
approach allowed researchers to specify independent variables, including<br />
the formation of a heterotypic interface <strong>and</strong> the ratio of cell populations<br />
at specific locations in their samples, something which was not possible<br />
using traditional r<strong>and</strong>om co-culture methods.<br />
Although many of the above findings have been made possible with<br />
the aid of micropatterning, they have also indicated the desirability of<br />
further investigation at smaller length scales (�100 nm) where most of the<br />
molecular mechanisms relevant to cell biology can be discovered. Thus,<br />
we describe relatively recent investigations that have used nanopatterned<br />
engineered surfaces in the next subsection.<br />
Nanopatterning<br />
To generate nanopatterns, electron beam lithography (EBL) is normally<br />
used since the spatial resolution of photolithography is limited<br />
by the diffraction of light. Rather than using a mask, EBL uses a focused<br />
electron beam to directly write patterns onto an electron beam sensitive<br />
resist. Since this is a serial writing method, i.e. tracks are written segment<br />
by segment, this technique can require a significant amount of time to<br />
write a single large area pattern <strong>and</strong> is thus very expensive. However, in a<br />
development similar to the soft lithography described earlier, nanopattern<br />
structures can also be imprinted onto a solid polymeric substrate (nanoimprinting<br />
lithography [NIL]), greatly reducing the cost [31, 32]; Figure 7.4).<br />
By combining NIL with self-assembled monolayer techniques, protein<br />
nanopatterns of dimension �100 nm have been produced [33, 34]. The<br />
combined capability of NIL <strong>and</strong> EBL for the generation of arbitrary nanopatterns<br />
on a wide range of materials has also led to the discovery of cell<br />
response to nanotopography, as discussed in later sections.<br />
Other methods for nanopatterning biological molecules include scanning<br />
probe lithography [35], self-assembly nanofabrication using block<br />
copolymer, <strong>and</strong> colloidal lithography [36]. A good review of these techniques<br />
is given by Gates et al. [37]. However, to generate a statistically<br />
meaningful cellular study, fine tailored adhesive nanopatterns have to<br />
cover a large area, preferably of the order of square centimetre. This<br />
imposes a big challenge for some of the serial writing methods, such as<br />
scan probe-associated lithography, although new developments in parallel<br />
writing using multiple tips might mitigate this barrier.<br />
As an example of an extension of the self-assembled block copolymer<br />
technique, recently, Spatz’s group have developed the ‘micelle diblock<br />
copolymer lithography’. This allows precise control of space between<br />
RGD lig<strong>and</strong>s at the length scale of 10–200 nm [38]. This strategy uses selfassembly<br />
of diblock polymer of polystyrene-block-poly(2-vinylpyridine)
204 7. MICRO/NANOENgINEERINg ANd AFM FOR CELLULAR SENSINg<br />
Hot<br />
embossing<br />
(a)<br />
(b) (e)<br />
Heat<br />
(c)<br />
(d)<br />
Heat<br />
into reverse micelles in toluene which form a uniform thin layer on a<br />
substrate. The cores of these micelles contain a metal precursor (HAuCl 4)<br />
that is turned into a regular Au nanoparticle upon oxygen plasma treatment<br />
of the film. Tailoring the ratios <strong>and</strong> components of the copolymer<br />
gives rise to a range of Au nanoparticles of 3–8 nm in diameter with<br />
spacing adjusted from 15 to 250 nm. By forming these fine Au nanoisl<strong>and</strong>s<br />
on a non-adhesive substrate <strong>and</strong> subsequent functionalision of the<br />
(f)<br />
(g)<br />
UV<br />
Flash<br />
imprint<br />
FIgurE 7.4 Schematic drawing of nanoimprint lithography (NIL). (a) A master with<br />
nanoscale topographic features is held above a sample coated in a layer of resist. (b) In hot<br />
embossing, the resist is a thermoplastic that can be softened with the application of heat.<br />
(c) The master is brought into contact with the heated resist <strong>and</strong> held under pressure.<br />
(d) The heat is removed, <strong>and</strong> the master is removed after the mask has cooled, leaving the<br />
topographic features embossed in the resist. (e) In flash imprint lithography, the resist is a<br />
viscous liquid or soft polymer that can be hardened through UV exposure. (f) The master is<br />
brought into contact with the resist <strong>and</strong> held under pressure whilst the sample is exposed<br />
to UV radiation. (g) The master is removed after the resist has hardened leaving the topographic<br />
features imprinted in the resist.
Au nanoparticles with a cyclic RGD-thiol, nanoisl<strong>and</strong>s of RGD-lig<strong>and</strong>s<br />
are created. This novel method has enabled a series of studies, demonstrating<br />
that cells can detect surface variations at the size of a typical protein<br />
complex (�100 nm) <strong>and</strong> are affected dramatically by small variations<br />
in lig<strong>and</strong>–cluster spacing (e.g. 58 <strong>and</strong> 73 nm) [39].<br />
Although nanopatterning of adhesive molecules is ultimately valuable<br />
to reveal molecular mechanisms underlying cellular processes, the<br />
reliability of the information obtained depends on the accuracy of the<br />
characterisation measurements at the nanoscale. In this context, it should<br />
be noted that topographic features as small as 10 nm can exert dramatic<br />
effects on a cell (more details in the next section). This type of observation<br />
is particularly important for protein nanopatterning studies based on the<br />
assembly of particle templates, as discussed in Spatz’s work.<br />
7.2.2 nanotopography<br />
7.2 ENgINEERINg THE ECM FOR PRObINg CELL SENSINg 205<br />
In contrast to the use of micro- <strong>and</strong> nanoscale patterns of (bio)chemical<br />
motifs described earlier, cell behaviour has also been extensively studied<br />
on patterns of micro- <strong>and</strong> nanoscale topographic features [40].<br />
Motivation for many of these studies comes from the observation that<br />
although the physical form of the ECM appears as a r<strong>and</strong>om meshwork,<br />
in fact, it contains enormously detailed nano- <strong>and</strong> microscale structures.<br />
For instance, a corrugation with a period of 68 nm has been recently<br />
observed on collagen fibres [41]. When looked at on the small scale, the<br />
ECM possesses nanopores, micro- <strong>and</strong> nanofibres, <strong>and</strong> peaks <strong>and</strong> depressions.<br />
Since the middle of the twentieth century, it has been increasingly<br />
recognised that cells react to microtopography. Early evidence has shown<br />
that cells adhere, align <strong>and</strong> move along fibres in the range of 30–100 �m<br />
[42–44]. In recent decades, the advance in micro- <strong>and</strong> nanofabrication has<br />
allowed intense investigations in this area <strong>and</strong> greatly push forward our<br />
underst<strong>and</strong>ing.<br />
Since the early 1960s, Curtis <strong>and</strong> his colleagues have studied the reaction<br />
of a number of cell types with various microtopographic structures<br />
[40, 45]. Microgrooves with a wide combination of widths <strong>and</strong> depths<br />
have been studied, <strong>and</strong> it was found that cells react to both depth <strong>and</strong><br />
width. In general, on deep <strong>and</strong> narrow grooves, cells tend to bridge<br />
between grooves, whilst on shallow grooves, they often follow the surface,<br />
although detailed reactions are cell type dependent [46]. Substantial<br />
evidence has been obtained, showing that adhesion or interaction with<br />
microscale topographic structures induces changes in cell cytoskeletal<br />
organisation, apoptosis (programmed cell death), macrophase activation<br />
(causing inflammatory reactions) <strong>and</strong> gene expression [47, 48]. The<br />
observed phenomena clearly demonstrated that microtopography influences<br />
cell development, <strong>and</strong> thus triggered researchers to pose questions
206 7. MICRO/NANOENgINEERINg ANd AFM FOR CELLULAR SENSINg<br />
about how cells sense the nanoworld, if indeed they can. For example,<br />
most of the commonly used cell culture dishes have a certain surface<br />
roughness, does this matter?<br />
R<strong>and</strong>om or Semi-R<strong>and</strong>om Defined Nanostructures<br />
The use of spontaneous phase separation of blended polymers or copolymers<br />
provides an efficient way to produce different nanoscale features<br />
of controllable depths over a large area [49, 50]. By spin coating blends<br />
of polystyrene (PS) <strong>and</strong> poly(4-bromostyrene) (PBrS) onto silicon wafers<br />
<strong>and</strong> varying the ratio of polymers <strong>and</strong> their concentrations, nanoisl<strong>and</strong>s<br />
of heights between 13 <strong>and</strong> 95 nm were produced (Figure 7.5) <strong>and</strong> tested<br />
using fibroblast <strong>and</strong> endothelial cell cultures [51]. It was found that an<br />
enhanced cell response was observed on the nanostructured isl<strong>and</strong>s<br />
compared to the planar PS control: 13 nm height isl<strong>and</strong>s increased cell<br />
spreading <strong>and</strong> proliferation as well as a broad up-regulation of many<br />
genes involved in proliferation <strong>and</strong> matrix synthesis [52]. However, 95 nm<br />
high isl<strong>and</strong>s reduce cell spreading <strong>and</strong> proliferation [51]. Thus, by altering<br />
only the height of nanoscale features, a different cell response could be<br />
induced, demonstrating the significant roles that nanotopography might<br />
exert on cells.<br />
Although polymer demixing can reliably control the z-depth of nanofeatures,<br />
there can be less control over the lateral dimensions. In addition, for<br />
many systems a possible effect from the differences in the chemistry of two<br />
FIgurE 7.5 AFM picture of an example of nanoisl<strong>and</strong>s made by polymer demixing<br />
process. Image courtesy of Drs. Matthew J. Dalby <strong>and</strong> Stanley Affrossman.
7.2 ENgINEERINg THE ECM FOR PRObINg CELL SENSINg 207<br />
polymers cannot be completely ruled out. Consequently, an alternative way<br />
for fast <strong>and</strong> low cost of fabrication of nanoscale topographic features has<br />
been sought through the use of colloidal lithography. Monodispersed <strong>and</strong><br />
nanosized colloids made using wet chemistry techniques are commercially<br />
available. They can self-assemble into a monolayer on a substrate, with the<br />
spacing between each tailored by their surface charge or functional linker<br />
groups. The resulting colloid assembly is effectively a ‘photoresist’ pattern<br />
whose lateral dimensions are determined by the colloid size <strong>and</strong> spacing.<br />
Using this approach, nanocolumns of 160 nm height, 100 nm in diameter<br />
<strong>and</strong> 230 nm spacing have been produced in a bulk poly(methyl methacrylate)<br />
(PMMA) polymer [53, 54]. In comparison to the non-structured control,<br />
nanocolumns reduced focal adhesions of cells, but significantly increased<br />
density of filopodia formation. As discussed earlier, filopodia formation is<br />
associated with focal complexes, indicating the involvement of nanotopographic<br />
features on integrin cluster formation.<br />
The fact that many of the ECM proteins are present as nanofibres is<br />
driving intensive research on engineered nanofibres as replacements for<br />
ECM components. Carbon nanofibre compactions have been investigated<br />
for osteoblast culture with potential application as orthopedic/dental<br />
implants [55]. When compared with using conventional carbon fibres<br />
(diameter �100 nm), osteoblasts proliferate faster <strong>and</strong> deposit more extracellular<br />
calcium (indicating osteoblastic bone formation) on the carbon<br />
nanofibre compactions (diameter �100 nm). In other studies, synthetic<br />
<strong>and</strong> natural polymeric nanofibres have also long been regarded as promising<br />
analogues of the ECM. Here, a vast selection of well-established<br />
methods can be used to tailor their chemistries to match those found in<br />
the native ECM. To produce the polymeric nanofibres themselves, electrospining<br />
has become perhaps the simplest <strong>and</strong> most efficient technique for<br />
producing materials which can be assembled into 2D <strong>and</strong> 3D non-woven<br />
fibrous mesh (see review by Pham) [56]. Interestingly, cell culture on these<br />
nanofibre matrixes demonstrated better attachment <strong>and</strong> increased proliferation<br />
compared to that on substrates made from larger size fibres.<br />
Nanofibre meshes have also been found to stimulate cells to develop<br />
phenotypical behaviour [57, 58]. For example, NIH 3T3 fibroblasts <strong>and</strong><br />
normal rat kidney cells grown on a polyamide nanofibre matrix displayed<br />
in vivo-like morphology <strong>and</strong> breast epithelial cells on the same matrix<br />
underwent morphogenesis into multicellular spheroids [58].<br />
All the above-mentioned examples provide evidence which suggest that<br />
nanotopography has a significant influence on cell adhesion, cytoskeletal<br />
organisation <strong>and</strong> morphogenesis. However, the mechanisms involved are<br />
poorly understood: We do not know whether the less well-defined lateral<br />
dimensions of nanoisl<strong>and</strong>s made by either polymer mixing or colloidal<br />
lithography play any significant role in cellular behaviour; the nanofibre<br />
matrix may provide a large surface to volume ratio structure that could
208 7. MICRO/NANOENgINEERINg ANd AFM FOR CELLULAR SENSINg<br />
possibly entrap significant amounts of ECM proteins from the culture<br />
medium; <strong>and</strong> it is not clear as to what extent the chemistry <strong>and</strong> texture of<br />
the nanofibre matrix (or the nanoisl<strong>and</strong>s) interact with cells. As a r<strong>and</strong>om<br />
nanophase meshwork with poorly defined ‘nanotopography’ features, it is<br />
difficult to isolate the many parameters which might affect cells.<br />
Precisely Defined Nanostructures<br />
E-beam lithography <strong>and</strong> associated techniques can produce precisely<br />
defined nanostructures, so that individual variables can be investigated<br />
thoroughly. Features as small as 3 nm can be reliably fabricated on a substrate<br />
[59] <strong>and</strong> using modern high throughput machines, sufficiently large<br />
areas can be written for cellular studies. Through combination of EBL<br />
<strong>and</strong> NIL techniques, arbitrary nanopatterns can be replicated in thermoplastic<br />
polymers. Using this method, a large number of replicates having<br />
identical patterns of designed nanofeatures have been produced on polylactide,<br />
polycarbonate, PMMA, <strong>and</strong> polycaprolactone [60–62]. This mass<br />
production using different materials has enabled direct comparison of the<br />
influence of substrates having different surface chemistries but the same<br />
nanotopography on cell behaviour for a number of cell lines.<br />
The use of arbitrary nanopatterns provides a very flexible <strong>and</strong> systematic<br />
way to explore the interactions between cells <strong>and</strong> the nanoworld, e.g.<br />
when highly regular arrays of nanopit structures with pit diameters of 35,<br />
75 <strong>and</strong> 120 nm were used for fibroblast growth. Within this range of subtle<br />
variation in pit size, it was found that cell spreading reduced <strong>and</strong> there<br />
was less apparent stress fibre formation [60]. Following on from this study,<br />
EBL was used to create different levels of disorders in the nanopatterns.<br />
Using these levels it has been found that human mesenchymal stem cells<br />
(MSCs) were prompted to produce more bone mineral when there was<br />
a certain degree of disorder to the array patterns of nanopits (Figure 7.6;<br />
[61]). However, highly ordered nanopits resulted in low to negligible cellular<br />
adhesion <strong>and</strong> osteroblast differentiation. This discovery suggested<br />
that nanotopography might be an efficient method to guide MSC cells to<br />
be used in regenerative medicine <strong>and</strong> tissue engineering devices, since at<br />
present, many examples of failed bone implants have been found to be<br />
associated with encapsulation by soft tissue without direct bone bonding.<br />
Clearly, substantial studies have advanced our underst<strong>and</strong>ing of the<br />
influence of topography on cellular processes. However, it is still to be<br />
resolved as to how cells detect <strong>and</strong> respond to these nanofeatures. Much<br />
evidence has shown that nanotopography influences cellular cytoskeleton<br />
formation, <strong>and</strong> thus is likely to modulate membrane receptor organisation,<br />
integrin cluster formation, <strong>and</strong> intracellular signalling – all of which are<br />
related to focal adhesion formation. It is not clear as to whether the mechanosensitivity<br />
of cells to physical topography features employs the same<br />
machinery that cells use to sense changes in surface chemistry (e.g. adhesive<br />
patterns). New investigations have started to underst<strong>and</strong> the changes
7.2 ENgINEERINg THE ECM FOR PRObINg CELL SENSINg 209<br />
Ordered pattern<br />
(a) (d)<br />
(b)<br />
200 μm (e)<br />
(c)<br />
1 μm<br />
200 μm<br />
in signalling <strong>and</strong> genetic analysis. With ever closer collaboration between<br />
biologists <strong>and</strong> engineers <strong>and</strong> the availability of advanced tools for nanoscale<br />
measurements, such as AFM, a greater underst<strong>and</strong>ing of the underlying<br />
mechanisms of cellular interactions will be possible. This will lead to the<br />
development of more effective methods for regenerative medicine.<br />
7.2.3 nanoscale Measurement: Challenges <strong>and</strong><br />
opportunities for AFM<br />
It might have not been possible to discover many of the findings<br />
related to nanotopography-induced cellular reactions without the AFM.<br />
AFM imaging permits reliable <strong>and</strong> routine characterisation of nanotopographic<br />
features with high resolution particularly in the z-direction,<br />
(f)<br />
Disordered pattern<br />
FIgurE 7.6 The effect of nanotopography on cell differentiation. SEM images of nanotopographies<br />
fabricated by EBL (a <strong>and</strong> d). The nanopits (120 nm diameter, 100 nm deep)<br />
arranged in a highly ordered square (a) <strong>and</strong> with each pit r<strong>and</strong>omly displaced from the<br />
square pattern (�50 nm from the true centre) (d). The disordered nanostructures stimulate<br />
the human MSCs to express the bone-specific ECM protein osteopontin, as shown in<br />
(e <strong>and</strong> f) (arrows) in contrast to no effect seen with the highly ordered pits (b <strong>and</strong> c). Figure<br />
reprinted from Dalby et al. [61] with permission.
210 7. MICRO/NANOENgINEERINg ANd AFM FOR CELLULAR SENSINg<br />
facilitating the discovery of the significant effects on cell behaviour with<br />
subtle variations in the height of the nanoisl<strong>and</strong>s (13–95 nm) as described<br />
earlier (Figure 7.5). It has also long been desirable to image both<br />
nanoscale functional cellular components <strong>and</strong> nanostructures simultaneously.<br />
In the past, biologists have employed SEM <strong>and</strong> immunostaining<br />
transmission electron microscopy (TEM) approaches to observe functional<br />
proteins (such as vinculin). However, these approaches require<br />
many steps of sample preparation, including cell fixing, immunostaining,<br />
sample drying, <strong>and</strong> thus many features can be disguised. Although AFM<br />
has been a powerful tool in the study of isolated biomolecular systems<br />
<strong>and</strong> their interactions, only more recently have nanoscale observations of<br />
living cells been achieved. The rapid expansion of the AFM approach to<br />
living cell studies has just started.<br />
Many biological processes involve forces, <strong>and</strong> this is very much the case<br />
for cells adhering to the ECM. However, quantification of these forces is<br />
not straightforward, because a whole cell produces only small forces in<br />
the nanonewton range. Furthermore, the dimensions of functional components<br />
of a cell range from subnanometre (nucleic acid) to micrometre<br />
(whole cells), <strong>and</strong> measurements can be complicated by the whole cell<br />
structure dynamically remodelling during cell activities [14]. AFM with<br />
the ability to measure forces as small as piconewton <strong>and</strong> distances �1 nm<br />
demonstrates great flexibility <strong>and</strong> versatility for investigating the mechano-<br />
physical events occurring during biological interactions ranging from<br />
those of a single nucleic acid (deoxyribonucleic acid [DNA]) to those associated<br />
with whole intact cells.<br />
AFM in conjunction with the colloidal probe techniques has further<br />
broadened its applications. Here, there are vast combinations in the choice,<br />
designs <strong>and</strong> functionalisation of probes, allowing a range of studies from a<br />
single biomolecular interaction to cell or polymer mechanics. The discovery<br />
that a cell exerts force on a substrate, as mentioned earlier, has initiated<br />
significant efforts to investigate the role of mechanical properties in<br />
cell activities. In this context, the AFM microsphere indentation technique<br />
has provided an assessment tool with high sensitivity <strong>and</strong> microscale resolution<br />
that can perform well-defined investigations into cell interactions<br />
with substrates of different elasticity. In pioneering studies in this field it<br />
has been found that cell growth, differentiation, spreading <strong>and</strong> migration<br />
are all regulated by the elasticity of the substrates [63, 64].<br />
As AFM is a surface-based technique with high resolution, the integration<br />
of AFM with optical microscopes is essential for the investigation of<br />
intracellular events during scanning. There have been long st<strong>and</strong>ing questions<br />
about the dynamics of structural adaptations of a cell in the context<br />
of mechanotransduction [65]. For example, how do cells use their cytoskeleton<br />
conformation to transduce a mechanical stimulus to the nucleus <strong>and</strong><br />
induce genomic variations? Frequently, there are also many requirements
for in situ monitoring of changes in intracellular signalling in combination<br />
with cellular physical variations as a result of the presence of drugs. These<br />
are most readily revealed by optical fluorescence assays. Consequently, in<br />
the last few years, instrument manufacturers have developed a number<br />
of combined AFM-optical microscope platforms that are now starting to<br />
benefit studies in the fields of bioengineering, cell engineering <strong>and</strong> basic<br />
cell biology. In particular, configurations involving confocal microscopy<br />
[66] <strong>and</strong> total internal reflection fluorescence microscopy [67] have already<br />
demonstrated their power in in situ living cell studies.<br />
Since the invention of the AFM in 1986, we have witnessed significant<br />
growth of the employment of AFM to probe biological systems. In the<br />
next section we focus on recent developments associated with intact cell<br />
measurement.<br />
7.3 AFM In CEll MEASurEMEnt<br />
The unique advantages of AFM in cell measurement lie in its capability<br />
of being able to simultaneously (1) image cell topology under nearphysiological<br />
conditions, (2) measure mechanical properties of living<br />
cells, <strong>and</strong> (3) monitor functional cellular components <strong>and</strong> intracellular<br />
processes in conjunction with optical microscopy. This section will demonstrate<br />
the great potential of AFM for investigating the interaction of<br />
cells with their environment.<br />
7.3.1 AFM Imaging of Cells<br />
7.3 AFM IN CELL MEASUREMENT 211<br />
In the early days of AFM imaging of cells, the main restriction was<br />
the limited scan size of the instruments [68], due to cells being relatively<br />
large structures. As soon as the first instruments with scan ranges of several<br />
micrometres were developed, they were deployed to image cells [69].<br />
Today, modern bio-AFM instruments integrate an AFM platform onto the<br />
stage of a conventional inverted optical microscope, thus enabling easy<br />
positioning of AFM tip over a particular region of cells (Figure 7.7(a)).<br />
This development has been very useful in identifying regions of interest<br />
in cell morphology <strong>and</strong> where cells react to the microstructured substrate<br />
(Figure 7.7(b)). In addition, the optical imaging also permits simultaneous<br />
monitoring of the lateral morphology of cells <strong>and</strong>/or intracellular<br />
signalling during the investigation by AFM [70]. Through a ‘direct overlay’<br />
technique, it is possible to integrate AFM with optical images <strong>and</strong><br />
use the optical image to guide AFM operation. This enables correlation of<br />
the biophysical <strong>and</strong> biochemical functionalities of a cell. In reality, the different<br />
operating principles of AFM <strong>and</strong> optical microscopy can result in
212 7. MICRO/NANOENgINEERINg ANd AFM FOR CELLULAR SENSINg<br />
(a) (b)<br />
FIgurE 7.7 Optical images of an AFM cantilever positioned over a cell on (a) a planar<br />
substrate <strong>and</strong> (b) a structured substrate.<br />
an inaccurate overlay of the two images; however, this can be overcome<br />
by the use of registration methods devised by the manufacturers [71].<br />
In many studies, fixed cells have been used to preserve cell morphology<br />
<strong>and</strong> maximise the imaging resolution with functional intracellular<br />
structures being revealed by staining. AFM <strong>and</strong> fluorescence images of<br />
an osteoblast cell cultured on a nanopatterned polycarbonate substrate<br />
prepared by EBL <strong>and</strong> NIL as described in an earlier section are shown in<br />
Figure 7.8. Although fluorescence images of the F-actin (light colour in<br />
Fig. 7.8(a)) <strong>and</strong> tubulin (light colour in Fig. 7.8(b)) elements within a cell<br />
protrusion already suggest a well-developed cytoskeleton structure, the<br />
AFM images reveal incredible details of the cell structure (Figure 7.8(c)–<br />
(f)). Both the well-spread straight actin stress fibres <strong>and</strong> the more curved<br />
tubulin network are visible. Here, the AFM image measures the height<br />
of the cell structures with a resolution that could not be obtained using<br />
techniques such as confocal microscopy. As is common practice with biological<br />
cell measurements, AFM data are recorded in two image channels:<br />
the height image shows an overall topography <strong>and</strong> the error signal image<br />
highlights nanometre-scale surface topography, revealing well-defined<br />
variations in the structure. As a cautionary note, it should be noted that<br />
although the fixing procedure eases AFM imaging, it can also cause damage<br />
<strong>and</strong> alteration of delicate cellular structures [72].<br />
Living cells, although more relevant to studies of most biological events,<br />
are one of the most challenging samples to image with AFM [73]. They<br />
are much softer than fixed cells <strong>and</strong> often react to the imaging process<br />
itself – sometimes retracting filopodia or excreting vesicles. Some cell types,<br />
e.g. fibroblasts are harder to image than others due to their restless nature
Slow (μm)<br />
40<br />
20<br />
7.3 AFM IN CELL MEASUREMENT 213<br />
(a) (b)<br />
0<br />
0 20<br />
Slow (μm)<br />
(e)<br />
Fast (μm)<br />
(c) (d)<br />
8<br />
6<br />
4<br />
2<br />
0<br />
40<br />
0 2 4 6 8<br />
Fast (μm)<br />
645.5 nm<br />
0 nm<br />
183.3 nm<br />
0 nm<br />
under the tip, <strong>and</strong> the height contours of some are so steep as to prove<br />
difficult for the AFM tip to follow topographically. Thus, there are many<br />
considerations to take into account when designing experiments using<br />
live cell imaging, such as operation modes, loading force, choice of tip<br />
Slow (μm)<br />
Slow (μm)<br />
(f)<br />
40<br />
20<br />
8<br />
6<br />
4<br />
2<br />
0<br />
0<br />
0 20<br />
Fast (μm)<br />
0 2 4 6 8<br />
Fast (μm)<br />
129.7 mV<br />
0 mV<br />
309.1 nm<br />
FIgurE 7.8 Simultaneous AFM <strong>and</strong> optical imaging of functional cell structures on a<br />
nanostructured substrate. Fluorescence images show an overview of the F-actin stress fibres<br />
(light colour in (a)) <strong>and</strong> tubulin (light colour in (b)) cytoskeleton protein distributed on a protrusion<br />
region (a <strong>and</strong> b). AFM height (c) <strong>and</strong> error images (d) revealing the detailed subcellular<br />
structures of the same region. The Directoverlay TM feature (JPK instrument Ltd) was<br />
employed for image overlay <strong>and</strong> positioning of the AFM tip, permitting precise location of a<br />
zoomed in region (identified in d). AFM images (e <strong>and</strong> f) of the zoomed in region, revealing<br />
the filopodial interacting with the underlying nanopatterned structures. AFM imaging was<br />
carried out in contact mode using a silicon nitride cantilever with spring constant 0.01 N m �1 .<br />
40<br />
0 mV
214 7. MICRO/NANOENgINEERINg ANd AFM FOR CELLULAR SENSINg<br />
<strong>and</strong> appropriate physiological conditions (medium temperature <strong>and</strong> pH).<br />
Although advances in cell immobilisation <strong>and</strong> AFM imaging modes now<br />
permit soft, live cells to be imaged, the process remains far from trivial [74].<br />
For living cell imaging, an environmental control chamber with suitable<br />
medium is essential to keep cells alive <strong>and</strong> viable. When imaging<br />
in contact mode, cantilevers with the lowest spring constants (0.003–<br />
0.06 N m �1 ) operating with a small loading force (�1 nN) are preferred.<br />
Following these basic considerations, it is possible to successfully image<br />
live cells. Figure 7.9(a)–(f) show height <strong>and</strong> error images of live 3T3 cells<br />
on a fibronectin-coated glass slide <strong>and</strong> a fibronectin-coated PDMS microgrooved<br />
substrate. A well-spread cell morphology with fibrous cytoskeleton<br />
structures can be seen clearly in Figure 7.9(a) in comparison to the<br />
restrained morphology of the cell on a flat PDMS substrate (Figure 7.9(c)<br />
<strong>and</strong> (d)). This can be further demonstrated by quantitative analysis of the<br />
cell height <strong>and</strong> its projecting area. It should be noted that cells tend to<br />
detach from the PDMS substrates during imaging. All of these observations<br />
demonstrate that 3T3 fibroblasts have much weaker adhesion to<br />
soft PDMS substrates. On microstructured PDMS structures, cells tend<br />
to position their nucleus in the trough of the channel, close to one edge<br />
<strong>and</strong> mostly attach filopodia to the upper surfaces (Figure 7.9(e) <strong>and</strong> (f)).<br />
This gives rise to a preferred cell alignment along the channel length, as<br />
discussed earlier. Thus, AFM imaging provides a quantitative <strong>and</strong> high<br />
resolution approach to studying the behaviour of living cells.<br />
7.3.2 Elasticity Measurement of living Cells<br />
On a soft sample surface, an approaching tip with controlled force will<br />
cause an indentation. Although this is often problematic for AFM imaging<br />
of living cells, it is a way to measure <strong>and</strong> map the elastic properties<br />
of the cell structure. For a force–distance measurement, the deflection of<br />
the cantilever is plotted as function of the z-separation of the probe <strong>and</strong><br />
the sample. On a stiff sample (e.g. glass), the deflection is proportional<br />
to the probe sample separation. However, on the soft sample, the movement<br />
of the tip will be less than that of the sample, <strong>and</strong> the difference is<br />
the indention of sample (Figure 7.10(a)).<br />
In order to be able to extract material properties from an AFM force–<br />
distance curve, we must interpret the deflection using an appropriate<br />
model. As a starting point, it is reasonable to consider an AFM force–<br />
distance curve as arising from a conical tip indenting a planar surface.<br />
This was first considered by Hertz in 1882 [75] (equation (7.1)) <strong>and</strong> later<br />
generalised by Sneddon [76].<br />
F � �<br />
2<br />
2 E<br />
⋅ ⋅<br />
π<br />
2<br />
1 � v<br />
( )<br />
⋅ ( �)<br />
tan<br />
(7.1)
Slow [μm]<br />
(a)<br />
Slow [μm]<br />
(c)<br />
Slow [μm]<br />
(e)<br />
100<br />
50<br />
0<br />
0 50<br />
Fast [μm]<br />
80<br />
60<br />
40<br />
20<br />
0<br />
0 50<br />
100<br />
20<br />
Fast [μm]<br />
0<br />
0 20<br />
Fast [μm]<br />
40<br />
7.3 AFM IN CELL MEASUREMENT 215<br />
718 nm<br />
0 nm<br />
2.774 μm<br />
0 μm<br />
7.749 nm<br />
0 μm<br />
This model has subsequently been applied to AFM data by many<br />
researchers [77, 78] <strong>and</strong> modified to account for a number of tip geometries<br />
[79]. In order to use the Hertz model to calculate the Young’s<br />
modulus (E) of a sample, several experimental parameters have to<br />
be known: the applied force (F), indentation depth (�), semi-opening<br />
Slow [μm]<br />
(b)<br />
Slow [μm]<br />
Slow [μm]<br />
(f)<br />
100<br />
50<br />
0<br />
0 50<br />
Fast [μm]<br />
80<br />
60<br />
40<br />
20<br />
0<br />
0 50<br />
40<br />
20<br />
Fast [μm]<br />
0<br />
0 20 40<br />
Fast [μm]<br />
428.7 mV<br />
0 mV<br />
2.038 V<br />
0 V<br />
1.375 V<br />
FIgurE 7.9 AFM images of living cells on different substrates, glass (a <strong>and</strong> b), flat<br />
PDMS (c <strong>and</strong> d) <strong>and</strong> a microgrooved PDMS substrate (e <strong>and</strong> f). The grooves in (e) <strong>and</strong> (f)<br />
have a 12.5 �m period <strong>and</strong> are 1 �m deep. All the substrates were coated with fibronectin.<br />
(d)<br />
0 V
216 7. MICRO/NANOENgINEERINg ANd AFM FOR CELLULAR SENSINg<br />
(a)<br />
On glass<br />
On nucleus<br />
On filopodia<br />
0.0<br />
–1400 –1200 –1000 –800 –600 –400 –200 0 200<br />
Piezo-displacement (nm)<br />
(b)<br />
angle of the tip (�, only for a conical tip) <strong>and</strong> Poisson’s ratio (v, normally<br />
assumed to be 0.5 as cells are virtually incompressible). Young’s<br />
modulus can then be calculated by taking � <strong>and</strong> F at a single point or<br />
by fitting to the contact region of a force distance (F–�) curve. In the<br />
single-point method, a significant uncertainty comes from the accuracy<br />
of determining the point where the tip first contacts the surface with zero<br />
force (contact point). This can be extremely difficult, <strong>and</strong> so it is normally<br />
recommended to fit the data in the contact region.<br />
δ<br />
d<br />
0.4<br />
0.3<br />
0.2<br />
0.1<br />
δ<br />
Force (nN)<br />
δ<br />
Contact point<br />
δ + d<br />
FIgurE 7.10 Schematic diagram showing indentation of a soft substrate by an AFM tip<br />
<strong>and</strong> cantilever (a). Contact point between the tip <strong>and</strong> substrate with no load or indentation<br />
of the sample (left image). Tip indenting the sample by a distance �, with a load caused by<br />
the cantilever deflection d, the total displacement of the tip–sample separation is given by<br />
� � d (right image). Examples of force curves generated on a solid glass substrate (no indentation),<br />
a cell nucleus region <strong>and</strong> a filopodia region (b). Note the heterogeneity of the cell.<br />
The indentation distance into the sample (�) can be corrected by subtracting the cantilever<br />
deflection (d) from the piezo-displacement.
7.4 CONCLUSIONS 217<br />
In using simple Hertz-based models, it is important to realise that they<br />
may be valid only for small indentations, in the region of 5–10% of the total<br />
sample depth. In the case of cells, this will result in the first 200–500 nm<br />
of indentation giving a valid fit, but at deeper indentations the underlying<br />
substrate will start to influence the data. To overcome this problem, there<br />
has been some work on models that can accommodate indentation deeper<br />
than 10% of the total sample thickness [80], although currently they are not<br />
as well established as the conventional Hertz model.<br />
Other considerations that should be taken into account when measuring<br />
the mechanical properties of cells include the inhomogeneity <strong>and</strong> non-elastic<br />
nature of a cell. For example, the indentation of cell nucleus <strong>and</strong> filopodia<br />
is quite different as shown in Figure 7.10(b). Although it is assumed<br />
that the cell is truly elastic for the purpose of data analysis, some of the<br />
energy delivered during indentation will be dissipated due to the viscous<br />
<strong>and</strong> slightly plastic nature of a cell. This effect can be seen in the velocity-<br />
dependent hysteresis that can be observed in some experiments [78, 81].<br />
Measurements may also vary between cells, or even on the same cell due<br />
to the internal components of cell moving beneath the indenting tip.<br />
Since Tao et al. first utilised AFM for quantitative measurement of local<br />
elastic properties of a biological sample (cow tibia) [82], there has been<br />
significant growth in the use of AFM for elasticity measurements of living<br />
cells [83–86]. It has been found that the elasticity (or stiffness) of different<br />
cell types can vary from 0.1 to 40 kPa [87], <strong>and</strong> cell elasticity might<br />
function as a quantitative indicator during cell differentiation. Many studies<br />
have also shown that cancer cells are substantially softer than normal<br />
cells [88, 89], indicating that quantitative analysis of mechanical properties<br />
could be used to differentiate between cancerous <strong>and</strong> normal cells<br />
with similar appearances. The quantitative information of local cell elasticity<br />
in nanoscale spatial resolution has also provided valuable insights<br />
into cellular processes such as cell spreading [90] <strong>and</strong> cell migration [91].<br />
Monitoring the change in elasticity while treating cells with drugs that<br />
disrupt specific cytoskeletal structures (i.e. F-actin, tubulin <strong>and</strong> intermediate<br />
filaments) reveals that the actin network mainly determines the elastic<br />
properties of living cells [66]. Apart from the continuous contribution to<br />
underst<strong>and</strong>ing fundamental cellular mechanisms, these developments are<br />
also promising for drug testing studies [92].<br />
7.4 ConCluSIonS<br />
Recent developments in micro- <strong>and</strong> nanoengineering have created<br />
many opportunities to investigate the complex processes in cellular biology.<br />
Engineered surfaces with micro- or nanoscale features, either chemical
218 7. MICRO/NANOENgINEERINg ANd AFM FOR CELLULAR SENSINg<br />
or physical in nature, have been reliably produced by various methods.<br />
Many of these structures have emerged from the novel combination of<br />
advances in micro- or nanofabrication, chemistry <strong>and</strong> biology. Surface<br />
patterning methods offer considerable flexibility in pattern design, lig<strong>and</strong><br />
specificity <strong>and</strong> density, <strong>and</strong> surface composition to investigate interactions<br />
of cells with chemical variations in the ECM. A vast range of engineered<br />
topographic features, from micro- to nanoscale, have been found to play a<br />
significant regulatory role in cellular proliferation, migration <strong>and</strong> differentiation.<br />
These studies provide significant insights into mechanistic questions<br />
of how cells are able to sense, integrate <strong>and</strong> respond to collective<br />
chemical <strong>and</strong> mechanical signals. In addition, they also demonstrate that<br />
engineered environments can regulate cell functions <strong>and</strong> fate, <strong>and</strong> thus<br />
open new opportunities of developing tailored biosubstrates for specific<br />
tissue cell types, although this is still at its early stage.<br />
Optical microscopes <strong>and</strong> SEM have been essential tools for many of<br />
these studies; however, they are subject to many constraints in exploring<br />
the molecular mechanisms at subcellular or cellular level, which are essential<br />
for the control of the biological pathway. AFM, having been established<br />
as an important tool in development of nanopatterned surfaces,<br />
has provided increased momentum to living system studies. Nanometre<br />
spatial imaging <strong>and</strong> quantitative measurement of mechanical properties<br />
of functional components of a living cell have been reliably achieved.<br />
Long st<strong>and</strong>ing questions in the nanoworld, such as whether it is nanoscale<br />
topographic features or chemical patterns that predominate in inducing<br />
cellular reactions [40], may be answerable. The combination of AFM<br />
with existing optical techniques has already shown itself to be a powerful<br />
tool, <strong>and</strong> further insights will be gained into in vitro cell differentiation <strong>and</strong><br />
disease diagnostics. Together with the engineered environments that can<br />
be employed to guide cell function <strong>and</strong> fate, it might be possible to move<br />
towards controlling biological pathways in cell culture, <strong>and</strong> so explore new<br />
medicines <strong>and</strong> therapies for human health.<br />
There are still many open challenges in cell biology <strong>and</strong> in physiological<br />
<strong>and</strong> pathological dysfunctions. The convergence of micro- or nanoengineering,<br />
AFM <strong>and</strong> interfacial chemistry with cell biology tools might<br />
provide new opportunities to address these challenges. For example, a better<br />
underst<strong>and</strong>ing of the complex processes associated with a whole intact<br />
cell interacting with individual molecules when cells are in contact with a<br />
surface will facilitate the diagnostics of arthoroplastic failures <strong>and</strong> improve<br />
existing implants. A more ambitious approach is to regenerate failed tissue<br />
outside a body; this requires an engineered ECM with the functionalities<br />
that are found in a body. Although there have been intensive efforts in<br />
these fields, researchers are still a long way from recreating an ECM with<br />
similar molecular architecture <strong>and</strong> functionalities of the ECM found in vivo.
AbbREvIATIONS ANd SyMbOLS 219<br />
Increasing collaborations between scientists <strong>and</strong> communities from various<br />
disciplines, including engineering, nanobiotechnology, materials, biology<br />
<strong>and</strong> clinical medicine, are essential to approach this goal.<br />
ACknowlEdgEMEntS<br />
H. Yin is supported by the Royal Society of Edinburgh as an RSE personal<br />
research fellow. G. McPhee thanks the EPSRC for his studentship.<br />
P. S. Dobson is supported by an RCUK Academic Research Fellowship. We<br />
would like to thank Dr Matthew Dalby for proof reading <strong>and</strong> offering constructive<br />
suggestions <strong>and</strong> samples. We also thank Drs Andrew Glide <strong>and</strong><br />
Mathis Riehle <strong>and</strong> Professor Jon Cooper for their support <strong>and</strong> discussions.<br />
AbbrEvIAtIonS And SyMbolS<br />
� Semi-opening angle of the tip °<br />
AFM Atomic force microscope<br />
DNA Deoxyribonucleic acid, a nucleic acid<br />
E Young’s modulus N m�2 EBL Electron beam lithography<br />
ECM Extracellular matrix<br />
F Applied force N<br />
MSCs Human mesenchymal stem cells<br />
NIL Nanoimprinting lithography<br />
PBrS Poly(4-bromostyrene)<br />
PDMS Poly(dimethysiloxne)<br />
PEG Poly(ethylene glycol)<br />
PMMA Poly(methyl methacrylate)<br />
PS Polystyrene<br />
RGD A peptide sequence consisting of<br />
Arg-Gly-Asp<br />
SAM Self-assembly monolayer<br />
SEM Scanning electron microscopy<br />
TEM Transmission electron microscopy<br />
V Poisson’s ratio, dimensionless<br />
� Indentation depth m
220 7. MICRO/NANOENgINEERINg ANd AFM FOR CELLULAR SENSINg<br />
references<br />
[1] B.M. Gumbiner, Cell adhesion: the molecular basis of tissue architecture <strong>and</strong> morphogenesis,<br />
Cell 84 (3) (1996) 345–357.<br />
[2] K.P. Campbell, Muscular-dystrophies – loss of cytoskeleton extracellular-matrix linkage,<br />
Cell 80 (5) (1995) 675–679.<br />
[3] A. Lukes, S. Mun-Bryce, M. Lukes, G.A. Rosenberg, Extracellular matrix degradation<br />
by metalloproteinases <strong>and</strong> central nervous system diseases, Mol. Neurobiol. 19 (3)<br />
(1999) 267–284.<br />
[4] R.H. Kramer, X.D. Shen, H. Zhou, Tumor cell invasion <strong>and</strong> survival in head <strong>and</strong> neck<br />
cancer, Cancer Metastasis Rev. 24 (1) (2005) 35–45.<br />
[5] T. Demuth, M.E. Berens, Molecular mechanisms of glioma cell migration <strong>and</strong> invasion,<br />
J. Neurooncol. 70 (2) (2004) 217–228.<br />
[6] M. Aumailley, B. Gayraud, Structure <strong>and</strong> biological activity of the extracellular matrix,<br />
J. Mol. Med. 76 (3–4) (1998) 253–265.<br />
[7] B. Alberts, A. Johnson, P. Walter, J. Lewis, Molecular Biology of the Cell, Garl<strong>and</strong><br />
Science, Taylor & Francis Group, New York, 2008.<br />
[8] A.N. Barclay, Concluding remarks <strong>and</strong> the challenge from the immune system,<br />
Faraday Discuss. (1998) 345–350.<br />
[9] E. Ruoslahti, M.D. Pierschbacher, New perspectives in cell adhesion: RGD <strong>and</strong><br />
integrins, Science 238 (4826) (1987) 491–497.<br />
[10] C.G. Galbraith, K.M. Yamada, M.P. Sheetz, The relationship between force <strong>and</strong> focal<br />
complex development, J. Cell Biol. 159 (4) (2002) 695–705.<br />
[11] D.A. Lauffenburger, A.F. Horwitz, Cell migration: a physically integrated molecular<br />
process, Cell 84 (3) (1996) 359–369.<br />
[12] K. Burridge, M. ChrzanowskaWodnicka, Focal adhesions, contractility, <strong>and</strong> signaling,<br />
Annu. Rev. Cell Dev. Biol. 12 (1996) 463–518.<br />
[13] K. Rottner, A. Hall, J.V. Small, Interplay between Rac <strong>and</strong> Rho in the control of substrate<br />
contact dynamics, Curr. Biol. 9 (12) (1999) 640–648.<br />
[14] N.Q. Balaban, U.S. Schwarz, D. Riveline, P. Goichberg, G. Tzur, I. Sabanay, D. Mahalu,<br />
S. Safran, A. Bershadsky, L. Addadi, B. Geiger, Force <strong>and</strong> focal adhesion assembly: a<br />
close relationship studied using elastic micropatterned substrates, Nat. Cell Biol. 3 (5)<br />
(2001) 466–472.<br />
[15] D. Riveline, E. Zamir, N.Q. Balaban, U.S. Schwarz, T. Ishizaki, S. Narumiya, Z. Kam,<br />
B. Geiger, A.D. Bershadsky, Focal contacts as mechanosensors: externally applied<br />
local mechanical force induces growth of focal contacts by an mDia1-dependent <strong>and</strong><br />
ROCK-independent mechanism, J. Cell Biol. 153 (6) (2001) 1175–1185.<br />
[16] N. Wang, J.P. Butler, D.E. Ingber, Mechanotransduction across the cell-surface <strong>and</strong><br />
through the cytoskeleton, Science 260 (5111) (1993) 1124–1127.<br />
[17] G. Maheshwari, G. Brown, D.A. Lauffenburger, A. Wells, L.G. Griffith, Cell adhesion<br />
<strong>and</strong> motility depend on nanoscale RGD clustering, J. Cell Sci. 113 (10) (2000)<br />
1677–1686.<br />
[18] C.Q. Lin, M.J. Bissell, Multifaceted regulation of cell – differentiation by extracellularmatrix,<br />
FASEB J. 7 (9) (1993) 737–743.<br />
[19] M. Madou, Fundamentals of Microfabrication: The Science of Miniturization, CRC<br />
Press, Boca Raton, FL, 1998.<br />
[20] C.D. Bain, J. Evall, G.M. Whitesides, Formation of monolayers by the coadsorption<br />
of thiols on gold – variation in the head group, tail group, <strong>and</strong> solvent, J. Am. Chem.<br />
Soc. 111 (18) (1989) 7155–7164.<br />
[21] G.T. HermansonBioconjugate Techniques, Academic press, San Diego, 2002.<br />
[22] C.D. Bain, G.M. Whitesides, Formation of monolayers by the coadsorption of thiols<br />
on gold – variation in the length of the alkyl chain, J. Am. Chem. Soc. 111 (18) (1989)<br />
7164–7175.
REFERENCES 221<br />
[23] Y.N. Xia, G.M. Whitesides, Soft lithography, Ann. Rev. Mater. Sci. 28 (1998) 153–184.<br />
[24] R.S. Kane, S. Takayama, E. Ostuni, D.E. Ingber, G.M. Whitesides, Patterning proteins<br />
<strong>and</strong> cells using soft lithography, Biomaterials 20 (23–24) (1999) 2363–2376.<br />
[25] A. Folch, M. Toner, Microengineering of cellular interactions, Ann. Rev. Biomed. Eng.<br />
2 (2000) 227–256.<br />
[26] K.L. Prime, G.M. Whitesides, Adsorption of proteins onto surfaces containing end –<br />
attached oligo(ethylene oxide) – a model system using self-assembled monolayers,<br />
J. Am. Chem. Soc. 115 (23) (1993) 10714–10721.<br />
[27] M. Mrksich, C.S. Chen, Y.N. Xia, L.E. Dike, D.E. Ingber, G.M. Whitesides, Controlling<br />
cell attachment on contoured surfaces with self-assembled monolayers of alkanethiolates<br />
on gold, Proc. Nat. Acad. Sci. U.S.A. 93 (20) (1996) 10775–10778.<br />
[28] E. Ostuni, R.G. Chapman, R.E. Holmlin, S. Takayama, G.M. Whitesides, A survey<br />
of structure–property relationships of surfaces that resist the adsorption of protein,<br />
Langmuir 17 (18) (2001) 5605–5620.<br />
[29] C.S. Chen, M. Mrksich, S. Huang, G.M. Whitesides, D.E. Ingber, Geometric control of<br />
cell life <strong>and</strong> death, Science 276 (5317) (1997) 1425–1428.<br />
[30] S.N. Bhatia, U.J. Balis, M.L. Yarmush, M. Toner, Effect of cell–cell interactions in preservation<br />
of cellular phenotype: cocultivation of hepatocytes <strong>and</strong> nonparenchymal cells,<br />
FASEB J. 13 (14) (1999) 1883–1900.<br />
[31] L.J. Guo, Nanoimprint lithography: methods <strong>and</strong> material requirements, Adv. Mater.<br />
19 (4) (2007) 495–513.<br />
[32] S.H. Ahn, L.J. Guo, High-speed roll-to-roll nanoimprint lithography on flexible plastic<br />
substrates, Adv. Mater. 20 (11) (2008) 2044–2049.<br />
[33] D. Falconnet, D. Pasqui, S. Park, R. Eckert, H. Schift, J. Gobrecht, R. Barbucci, M. Textor,<br />
A novel approach to produce protein nanopatterns by combining nanoimprint lithography<br />
<strong>and</strong> molecular self-assembly, Nano Lett. 4 (10) (2004) 1909–1914.<br />
[34] J.D. Hoff, L.J. Cheng, E. Meyhofer, L.J. Guo, A.J. Hunt, Nanoscale protein patterning<br />
by imprint lithography, Nano Lett. 4 (5) (2004) 853–857.<br />
[35] R.D. Piner, J. Zhu, F. Xu, S.H. Hong, C.A. Mirkin, “Dip-pen” nanolithography, Science<br />
283 (5402) (1999) 661–663.<br />
[36] X.M. Li, J. Huskens, D.N. Reinhoudt, Reactive self-assembled monolayers on flat <strong>and</strong><br />
nanoparticle surfaces, <strong>and</strong> their application in soft <strong>and</strong> scanning probe lithographic<br />
nanofabrication technologies, J. Mater. Chem. 14 (20) (2004) 2954–2971.<br />
[37] B.D. Gates, Q.B. Xu, M. Stewart, D. Ryan, C.G. Willson, G.M. Whitesides, New<br />
approaches to nanofabrication: molding, printing, <strong>and</strong> other techniques, Chem. Rev.<br />
105 (4) (2005) 1171–1196.<br />
[38] P.P. Girard, E.A. Cavalcanti-Adam, R. Kemkemer, J.P. Spatz, Cellular chemomechanics<br />
at interfaces: sensing, integration <strong>and</strong> response, Soft Matter 3 (3) (2007) 307–326.<br />
[39] M. Arnold, E.A. Cavalcanti-Adam, R. Glass, J. Blummel, W. Eck, M. Kantlehner,<br />
H. Kessler, J.P. Spatz, Activation of integrin function by nanopatterned adhesive interfaces,<br />
Chemphyschem 5 (3) (2004) 383–388.<br />
[40] A.S.G. Curtis, C.D. Wilkinson, Reactions of cells to topography, J. Biomater. Sci.<br />
Polym. Ed. 9 (12) (1998) 1313–1329.<br />
[41] D. Meller, K. Peters, K. Meller, Human cornea <strong>and</strong> sclera studied by atomic force<br />
microscopy, Cell Tissue Res. 288 (1) (1997) 111–118.<br />
[42] P. Weiss, Experiments on cell <strong>and</strong> axon orientation in vitro – the role of colloidal exudates<br />
in tissue organization, J. Exp. Zool. 100 (3) (1945) 353–386.<br />
[43] P. Weiss, Cell Contact, Int. Rev. Cytol. Surv. Cell Biol. 7 (1958) 391–423.<br />
[44] P. Weiss, B. Garber, Shape <strong>and</strong> movement of mesenchyme cells as functions of the<br />
physical structure of the medium – contributions to a quantitative morphology, Proc.<br />
Natl. Acad. Sci. U.S.A 38 (3) (1952) 264–280.<br />
[45] A.S.G. Curtis, M. Varde, Control of cell behavior – topological factors, J. Natl. Cancer<br />
Inst. 33 (1) (1964) 15.
222 7. MICRO/NANOENgINEERINg ANd AFM FOR CELLULAR SENSINg<br />
[46] A. Curtis, C. Wilkinson, Topographical control of cells, Biomaterials 18 (24) (1997)<br />
1573–1583.<br />
[47] P. Clark, P. Connolly, A.S.G. Curtis, J.A.T. Dow, C.D.W. Wilkinson, Topographical control<br />
of cell behavior. 1. Simple step cues, Development 99 (3) (1987) 439–448.<br />
[48] P. Clark, P. Connolly, A.S.G. Curtis, J.A.T. Dow, C.D.W. Wilkinson, Topographical control<br />
of cell behavior. 2. Multiple grooved substrate, Development 108 (4) (1990) 635–644.<br />
[49] S. Affrossman, G. Henn, S.A. Oneill, R.A. Pethrick, M. Stamm, Surface topography <strong>and</strong><br />
composition of deuterated polystyrene–poly(bromostyrene) blends, Macromolecules<br />
29 (14) (1996) 5010–5016.<br />
[50] N. Gadegaard, M.J. Dalby, M.O. Riehle, A.S.G. Curtis, S. Affrossman, Tubes with<br />
controllable internal nanotopography, Adv. Mater. 16 (20) (2004) 1857–1860.<br />
[51] M.J. Dalby, M.O. Riehle, H.J.H. Johnstone, S. Affrossman, A.S.G. Curtis, Polymerdemixed<br />
nanotopography: control of fibroblast spreading <strong>and</strong> proliferation, Tissue<br />
Eng. 8 (6) (2002) 1099–1108.<br />
[52] M.J. Dalby, S.J. Yarwood, M.O. Riehle, H.J.H. Johnstone, S. Affrossman, A.S.G. Curtsi,<br />
Increasing fibroblast response to materials using nanotopography: morphological <strong>and</strong><br />
genetic measurements of cell response to 13-nm-high polymer demixed isl<strong>and</strong>s, Exp.<br />
Cell Res. 276 (1) (2002) 1–9.<br />
[53] M.J. Dalby, M.O. Riehle, D.S. Sutherl<strong>and</strong>, H. Agheli, A.S.G. Curtis, Changes in fibroblast<br />
morphology in response to nano-columns produced by colloidal lithography,<br />
Biomaterials 25 (23) (2004) 5415–5422.<br />
[54] M.J. Dalby, M.O. Riehle, D.S. Sutherl<strong>and</strong>, H. Agheli, A.S.G. Curtis, Fibroblast response<br />
to a controlled nanoenvironment produced by colloidal lithography, J. Biomed. Mater.<br />
Res. A 69A (2) (2004) 314–322.<br />
[55] K.L. Elias, R.L. Price, T.J. Webster, Enhanced functions of osteoblasts on nanometer<br />
diameter carbon fibers, Biomaterials 23 (15) (2002) 3279–3287.<br />
[56] Q.P. Pham, U. Sharma, A.G. Mikos, Electrospinning of polymeric nanofibers for tissue<br />
engineering applications: a review, Tissue Eng. 12 (5) (2006) 1197–1211.<br />
[57] X.M. Mo, C.Y. Xu, M. Kotaki, S. Ramakrishna, Electrospun P(LLA-CL) nanofiber: a<br />
biomimetic extracellular matrix for smooth muscle cell <strong>and</strong> endothelial cell proliferation,<br />
Biomaterials 25 (10) (2004) 1883–1890.<br />
[58] M. Schindler, I. Ahmed, J. Kamal, A. Nur-E-Kamal, T.H. Grafe, H.Y. Chung, S. Meiners,<br />
A synthetic nanofibrillar matrix promotes in vivo-like organization <strong>and</strong> morphogenesis<br />
for cells in culture, Biomaterials 26 (28) (2005) 5624–5631.<br />
[59] D.R.S. Cumming, S. Thoms, S.P. Beaumont, J.M.R. Weaver, Fabrication of 3 nm wires<br />
using 100 keV electron beam lithography <strong>and</strong> poly(methyl methacrylate) resist, Appl.<br />
Phys. Lett. 68 (3) (1996) 322–324.<br />
[60] M.J. Dalby, N. Gadegaard, M.O. Riehle, C.D.W. Wilkinson, A.S.G. Curtis, Investigating<br />
filopodia sensing using arrays of defined nano-pits down to 35 nm diameter in size,<br />
Int. J. Biochem. Cell Biol. 36 (10) (2004) 2005–2015.<br />
[61] M.J. Dalby, N. Gadegaard, R. Tare, A. Andar, M.O. Riehle, P. Herzyk, C.D.W.<br />
Wilkinson, R.O.C. Oreffo, The control of human mesenchymal cell differentiation<br />
using nanoscale symmetry <strong>and</strong> disorder, Nat. Mater. 6 (12) (2007) 997–1003.<br />
[62] M.J. Dalby, M.J.P. Biggs, N. Gadegaard, G. Kalna, C.D.W. Wilkinson, A.S.G. Curtis,<br />
Nanotopographical stimulation of mechanotransduction <strong>and</strong> changes in interphase<br />
centromere positioning, J. Cell. Biochem. 100 (2) (2007) 326–338.<br />
[63] T. Yeung, P.C. Georges, L.A. Flanagan, B. Marg, M. Ortiz, M. Funaki, N. Zahir,<br />
W.Y. Ming, V. Weaver, P.A. Janmey, Effects of substrate stiffness on cell morphology,<br />
cytoskeletal structure, <strong>and</strong> adhesion, Cell Motil. Cytoskeleton 60 (1) (2005) 24–34.<br />
[64] A.J. Engler, S. Sen, H.L. Sweeney, D.E. Discher, Matrix elasticity directs stem cell lineage<br />
specification, Cell 126 (4) (2006) 677–689.<br />
[65] D.E. Ingber, I. Tensegrity, Cell structure <strong>and</strong> hierarchical systems biology, J. Cell Sci.<br />
116 (7) (2003) 1157–1173.
REFERENCES 223<br />
[66] G.T. Charras, M.A. Horton, Single cell mechanotransduction <strong>and</strong> its modulation analyzed<br />
by atomic force microscope indentation, Biophys. J. 82 (6) (2002) 2970–2981.<br />
[67] A.B. Mathur, G.A. Truskey, W.M. Reichert, Atomic force <strong>and</strong> total internal reflection<br />
fluorescence microscopy for the study of force transmission in endothelial cells,<br />
Biophys. J. 78 (4) (2000) 1725–1735.<br />
[68] H.G. Hansma, J.H. Hoh, Biomolecular imaging with the atomic-force microscope,<br />
Ann. Rev. Biophys. Biomol. Struct. 23 (1994) 115–139.<br />
[69] S.A.C. Gould, B. Drake, C.B. Prater, A.L. Weisenhorn, S. Manne, H.G. Hansma,<br />
P.K. Hansma, J. Massie, M. Longmire, V. Elings, B.D. Northern, B. Mukergee, C.M.<br />
Peterson, W. Stoeckenius, T.R. Albrecht, C.F. Quate, From atoms to integrated-circuit<br />
chips, blood-cells, <strong>and</strong> bacteria with the atomic force microscope, in: 4th International<br />
Conference on Scanning Tunneling Microscopy/Spectroscopy, Oarai, Japan, 1989.<br />
[70] Y.R. Silberberg, A.E. Pelling, G.E. Yakubov, W.R. Crum, D.J. Hawkes, M.A. Horton,<br />
Mitochondrial displacements in response to nanomechanical forces, J. Mol. Recognit.<br />
21 (1) (2008) 30–36.<br />
[71] JPK. Application notes. Available from: http://www.jpk.com/index.2.html.<br />
[72] S.H. Doak, D. Rogers, B. Jones, L. Francis, R.S. Conlan, C. Wright, High-resolution<br />
imaging using a novel atomic force microscope <strong>and</strong> confocal laser scanning microscope<br />
hybrid instrument: essential sample preparation aspects, Histochem. Cell Biol.<br />
130 (5) (2008) 909–916.<br />
[73] H.J. Butt, E.K. Wolff, S.A.C. Gould, B.D. Northern, C.M. Peterson, P.K. Hansma,<br />
Imaging cells with the atomic force microscope, J. Struct. Biol. 105 (1–3) (1990) 54–61.<br />
[74] Y.F. Dufrene, Nanoscale exploration of microbial surfaces using the atomic force<br />
microscope, Future Microbiol. 1 (4) (2006) 387–396.<br />
[75] H. Hertz, Uber die Beruhrug fester elastischer Korper, Journal fur die Reine und<br />
Angew<strong>and</strong>te Mathematik 92 (1882) 156–171.<br />
[76] I.N. Sneddon, The relation between load <strong>and</strong> penetration in the axisymmetric boussinesq<br />
problem for a punch of arbitrary profile, Int. J. Eng. Sci. 3 (1) (1965) 47–57.<br />
[77] M. Radmacher, Measuring the elastic properties of biological samples with the AFM,<br />
IEEE Eng. Med. Biol. Mag. 16 (2) (1997) 47–57.<br />
[78] M.J. Rosenbluth, W.A. Lam, D.A. Fletcher, Force microscopy of nonadherent cells: a<br />
comparison of leukemia cell deformability, Biophys. J. 90 (8) (2006) 2994–3003.<br />
[79] D.C. Lin, E.K. Dmitriadis, F. Horkay, Robust strategies for automated AFM force curve<br />
analysis-II: adhesion-influenced indentation of soft, elastic materials, J. Biomech. Eng.<br />
Trans. ASME 129 (6) (2007) 904–912.<br />
[80] D.C. Lin, F. Horkay, Nanomechanics of polymer gels <strong>and</strong> biological tissues: a critical<br />
review of analytical approaches in the Hertzian regime <strong>and</strong> beyond, Soft Matter 4 (4)<br />
(2008) 669–682.<br />
[81] Q.S. Li, G.Y.H. Lee, C.N. Ong, C.T. Lim, AFM indentation study of breast cancer cells,<br />
Biochem. Biophys. Res. Commun. 374 (4) (2008) 609–613.<br />
[82] N.J. Tao, S.M. Lindsay, S. Lees, Measuring the microelastic properties of biological –<br />
material, Biophys. J. 63 (4) (1992) 1165–1169.<br />
[83] J.H. Hoh, C.A. Schoenenberger, Surface-morphology <strong>and</strong> mechanical-properties of<br />
MDCK monolayers by atomic-force microscopy, J. Cell Sci. 107 (1994) 1105–1114.<br />
[84] J.A. Dvorak, E. Nagao, Kinetic analysis of the mitotic cycle of living vertebrate cells by<br />
atomic force microscopy, Exp. Cell Res. 242 (1) (1998) 69–74.<br />
[85] C. Rotsch, K. Jacobson, M. Radmacher, Dimensional <strong>and</strong> mechanical dynamics of<br />
active <strong>and</strong> stable edges in motile fibroblasts investigated by using atomic force microscopy,<br />
Proc. Natl. Acad. Sci. U.S.A. 96 (3) (1999) 921–926.<br />
[86] H.J. Butt, B. Cappella, M. Kappl, Force measurements with the atomic force microscope:<br />
technique, interpretation <strong>and</strong> applications, Surf. Sci. Rep. 59 (1–6) (2005) 1–152.<br />
[87] P.A. Janmey, C.A. McCulloch, Cell mechanics: integrating cell responses to mechanical<br />
stimuli, Ann. Rev. Biomed. Eng. 9 (2007) 1–34.
224 7. MICRO/NANOENgINEERINg ANd AFM FOR CELLULAR SENSINg<br />
[88] S.E. Cross, Y.S. Jin, J. Rao, J.K. Gimzewski, Nanomechanical analysis of cells from<br />
cancer patients, Nat. Nanotechnol. 2 (12) (2007) 780–783.<br />
[89] S. Suresh, Biomechanics <strong>and</strong> biophysics of cancer cells, Acta Biomater. 3 (4) (2007)<br />
413–438.<br />
[90] L. Soon, F. Braet, J. Condeelis, Moving in the right direction – nanoimaging in cancer<br />
cell motility <strong>and</strong> metastasis, Microsc. Res. Tech. 70 (3) (2007) 252–257.<br />
[91] A.A. Wagh, E. Roan, K.E. Chapman, L.P. Desai, D.A. Rendon, E.C. Eckstein,<br />
C.M. Waters, Localized elasticity measured in epithelial cells migrating at a wound<br />
edge using atomic force microscopy, Am. J. Physiol. Lung Cell. Mol. Physiol. 295 (1)<br />
(2008) L54–L60.<br />
[92] H. Oberleithner, C. Riethmuller, T. Ludwig, M. Hausberg, H. Schillers, Aldosterone<br />
remodels human endothelium, in: E.K. Hoffman (ed) International Symposium Cell<br />
Volume Control in Health <strong>and</strong> Disease, Blackwell Publishing, Copenhagen, Denmark,<br />
2005.
C H A P T E R<br />
8<br />
Atomic Force Microscopy<br />
<strong>and</strong> Polymers on Surfaces<br />
Vasileios Koutsos<br />
O U T L I N E<br />
8.1 Introduction 225<br />
8.2 Basic Concepts 227<br />
8.3 End-grafted Polymer Chains 229<br />
8.4 Diblock Copolymers Adsorbed on Surfaces 236<br />
8.5 Star-shaped Polymers Adsorbed on Surfaces 237<br />
8.6 Conclusions 240<br />
Acknowledgements 241<br />
List of Abbreviations 242<br />
List of Symbols 242<br />
References 242<br />
8.1 IntroDuCtIon<br />
The surfaces of materials are routinely modified with coatings to protect<br />
them in hostile conditions <strong>and</strong> to functionalise them for a variety of<br />
purposes. Polymer coatings play an important role in such modifications.<br />
Polymer synthetic chemists are nowadays able to produce a huge number<br />
of different macromolecules of controlled structure <strong>and</strong> composition<br />
(<strong>and</strong> consequently function) in large quantities <strong>and</strong> at a relatively low cost.<br />
Furthermore, polymers are usually processed readily <strong>and</strong> inexpensively.<br />
Atomic Force Microscopy in Process Engineering 225<br />
© 2009, Elsevier Ltd
226 8. ATOMIC FORCE MICROSCOPy ANd POLyMERS ON SURFACES<br />
Ultrathin polymer coatings, in particular, play a crucial role in many<br />
processes, ranging from decoration <strong>and</strong> protection against a variety of<br />
degradation agents (corrosion/chemical) [1] to microfabrication methodologies<br />
for microelectronics [2] <strong>and</strong> biomedical devices [3]. In all situations,<br />
at the polymer-solid interface, there is a layer of polymer chains<br />
that are in direct contact with the solid surface. The coating can be so<br />
thin that it actually consists of just one mono-macromolecular layer, i.e.<br />
a monolayer of polymer chains that are all in contact with the solid surface.<br />
Such monolayers are typically some nanometres thick (the thickness<br />
is of the order of the typical size of one polymer chain). In many<br />
cases, the solid surface is not fully covered with polymer chains, <strong>and</strong> we<br />
have the formation of a submonolayer, the structure of which can range<br />
from a complete layer occasionally disrupted by some isolated holes to<br />
a solid surface only partially covered by some isolated polymer isl<strong>and</strong>s.<br />
Depending on the molecular interactions <strong>and</strong> the number of chains at<br />
the surface (surface coverage), there is a variety of different continuous,<br />
semi-continuous <strong>and</strong> discontinuous nanopatterns <strong>and</strong> nanostructures<br />
that can be formed.<br />
Polymer monolayers <strong>and</strong> submonolayers on solid surfaces play an<br />
important role for many technological applications such as nanopatterning<br />
<strong>and</strong> surface modifications used in microelectronics [4], colloidal stability<br />
<strong>and</strong> flocculation [5], polymer reinforcement with nanofillers [6, 7],<br />
non-fouling biosurfaces [8], biocompatibility of medical implants [9],<br />
separations [10], microfluidics [11], adhesion, lubrication <strong>and</strong> friction<br />
modification [12]. Of particular importance are recent developments in<br />
the direction of achieving smart responsive surfaces to various external<br />
stimuli using polymer monolayers <strong>and</strong> nanostructures [13–15].<br />
The atomic force microscope (AFM) [16] provides a spatial <strong>and</strong> force<br />
resolution in the order of angstroms <strong>and</strong> tens of piconewtons, respectively;<br />
therefore, it is ideally suited to study the fine morphology <strong>and</strong><br />
physico-chemical properties of materials at the nanometre scale, directly<br />
in real space. This is of particular importance in the field of polymer-<br />
modified material surfaces. The traditional surface techniques lack the<br />
lateral resolution at the nanometre scale, which is necessary in order<br />
to analyse laterally inhomogeneous monolayers <strong>and</strong> submonolayers of<br />
polymers, macromolecular-size clusters <strong>and</strong> polymer nanostructures or<br />
nanopatterns. Furthermore, the AFM can operate under liquid conditions,<br />
providing a direct look of polymers anchored on surfaces in various<br />
solvent conditions. In many cases, solution-based processing techniques<br />
(e.g. dip-coating, spin-coating, droplet-evaporation, spraying) are employed<br />
<strong>and</strong> consequently, direct measurements within a good solvent (for the<br />
polymer) are of particular importance. AFM investigations within liquids<br />
can reveal <strong>and</strong> elucidate the physico-chemical phenomena governing<br />
the polymer monolayer behaviour during formation <strong>and</strong> processing.
8.2 BASIC CONCEPTS 227<br />
Fur-thermore, in many cases, mainly within biotechnology applications,<br />
the functional use of the polymer monolayer is under liquid (e.g. aqueous)<br />
conditions <strong>and</strong> in these cases AFM provides the opportunity for direct,<br />
real-space <strong>and</strong> real-time monitoring of the polymer layer during its functional<br />
role. In any case, it has to be noted that AFM studies of dry polymer<br />
monolayers are more numerous since they are easier to implement. To<br />
some extent, they can provide important snapshots of polymeric behaviour<br />
in the solution <strong>and</strong> can elucidate many aspects of the polymer behaviour.<br />
In this chapter, first, a concise <strong>and</strong> simplified version of the fundamental<br />
physical ideas of how polymer chains behave when they are in close<br />
proximity to surfaces is presented; the following part is dedicated to the<br />
main objective of this chapter, which is to demonstrate the potential, versatility<br />
<strong>and</strong> flexibility of the AFM technique to probe in a direct manner<br />
the nanoscale structural <strong>and</strong> physical properties of polymer monolayers<br />
<strong>and</strong> submonolayers. To this end, we use some selected examples from<br />
our own AFM studies. We show that the structural regimes depend on<br />
many factors such as surface interactions, polymer molecular weight,<br />
surface density, solvent conditions, chemical composition <strong>and</strong> molecular<br />
architecture. The structural properties of these ultrathin polymer films<br />
have a profound effect on various chemomechanical properties of the<br />
modified surfaces.<br />
In summary, we show in a direct manner the important role of the<br />
AFM as a very appropriate characterisation technique <strong>and</strong> tool for the<br />
investigation of materials surfaces that has been modified <strong>and</strong> processed<br />
by polymer monolayers <strong>and</strong> submonolayers.<br />
8.2 BASIC ConCEPtS<br />
A polymer chain can be attached on a surface by physical (e.g. van der<br />
Waals, electrostatic forces) <strong>and</strong> chemical (e.g. covalent) bonds. Its behaviour<br />
depends strongly on the characteristics of the attachment such as the<br />
number <strong>and</strong> position of attachment points (e.g. anchoring by its end or<br />
attachment points along the backbone of the chain) <strong>and</strong> bond strength,<br />
with chemical bonds being usually stronger than any physical attractive<br />
forces <strong>and</strong> hydrogen bonds (which are ranked as of moderate strength).<br />
One has to note that a large number of weak physical bonds can be a very<br />
efficient way of attachment <strong>and</strong> in this way a polymer chain can take a<br />
flat conformation (Figure 8.1). Nevertheless, the adsorption by chemical<br />
bonds (chemisorption) is considered irreversible, while adsorption by<br />
physical attractive forces (physisorption) is usually reversible under certain<br />
processing conditions.<br />
A polymer chain in good solvent conditions (e.g. polystyrene in toluene)<br />
is swollen since the polymer segments repel each other because of the
228 8. ATOMIC FORCE MICROSCOPy ANd POLyMERS ON SURFACES<br />
Tails<br />
Loops<br />
(a) (b) (c)<br />
FIgurE 8.1 (a) A polymer chain chemically attached by one of its ends to a solid surface<br />
‘swollen’ in a good solvent. (b) A polymer chain physisorbed on a solid surface by nonspecific<br />
(physical) interactions along its backbone (four contact points), forming three loops<br />
<strong>and</strong> two tails. (c) Diblock copolymer in a solvent attached on a solid surface by adsorption<br />
of one of its two blocks.<br />
repulsive excluded volume interactions between the monomers. If the<br />
solvent conditions change to bad (e.g. polystyrene in water or in dry state),<br />
then the polymer chain collapses into a dense globule as the monomer-<br />
monomer attractive forces prevail <strong>and</strong> the polymer tries to minimise its<br />
contact area with the unfavourable solvent (Figure 8.2a, b). If the polymer<br />
chain happens to be attached on the surface, the behaviour is similar<br />
(Figure 8.2c,d) but the final conformation is to be affected by the strength<br />
of monomer-surface interactions <strong>and</strong> location of the attachment points. An<br />
isolated end-grafted polymer chain with negligible interactions (apart from<br />
the end-grafting) with the solid surface resembles a mushroom (especially<br />
when it is swollen in good solvent conditions), <strong>and</strong> for such grafting densities,<br />
the polymer chains are within the ‘mushroom’ regime (Figures 8.1a<br />
<strong>and</strong> 8.2b,c). In contrast, if the polymer-surface interactions are high enough<br />
<strong>and</strong> involve very many polymer segments, the polymer takes a flat conformation,<br />
which resembles a pancake, signifying the ‘pancake’ regime.<br />
If the grafting density is high enough, the polymer chains start to<br />
overlap <strong>and</strong> interact with each other, <strong>and</strong> consequently the whole polymer<br />
layer can increase its thickness (compared with the dimensions of<br />
a single polymer chain in the corresponding solvent conditions). This<br />
effect is particularly dramatic in good solvent conditions since in this<br />
case the excluded volume interactions between the polymer segments<br />
are repulsive <strong>and</strong> the whole polymer layer is stretched away in the perpendicular<br />
to the solid surface direction, as depicted in Figure 8.3(a). For<br />
sufficiently high grafting densities, the polymer layer resembles a brush<br />
[17] <strong>and</strong> it is called a ‘polymer brush’. What will happen to the morphology<br />
of such polymer monolayers upon drying? Will the polymer chains<br />
collapse individually or in a homogeneous layer or even in aggregates<br />
(Figure 8.3)? Could one use such a process to form useful nanostructures<br />
<strong>and</strong> nanopatterns? Furthermore, what is the effect of polymer composition<br />
or polymer architecture <strong>and</strong> what are the nanomechanical properties<br />
of such polymer layers?
(a)<br />
Good solvent:<br />
swollen state<br />
8.3 ENd-gRAFTEd POLyMER CHAINS 229<br />
(c) (d)<br />
Bad solvent:<br />
collapsed state<br />
FIgurE 8.2 (a) A swollen polymer chain in the ‘bulk’ solution <strong>and</strong> (b) in a collapsed<br />
globular state upon the change of the solvent conditions from good to bad. (c,d) The same<br />
transition for an end-grafted polymer chain with negligible affinity for the solid surface<br />
(mushroom regime).<br />
In the following sections, we will be demonstrating the AFM capability<br />
to give answers to these <strong>and</strong> other similar questions using specific examples<br />
of chemisorbed or physisorbed polymer chains on solid surfaces from<br />
dilute polymer solutions. Owing to the characteristic nanometre-scale size<br />
of single chains <strong>and</strong> thickness of polymer monolayers, any inhomogeneity,<br />
aggregation, instabilitity <strong>and</strong> dewetting effects of such layers can be<br />
readily <strong>and</strong> directly probed by atomic force microscopy, which is based on<br />
piezoelectric transducers <strong>and</strong> sharp probing tips mounted on microfabricated<br />
cantilevers that scan the interrogated surface <strong>and</strong> are simultaneously<br />
monitored by the laser beam deflection technique while a feedback loop<br />
controls the distance between the probing tip <strong>and</strong> sample.<br />
8.3 EnD-grAFtED PoLymEr ChAInS<br />
An efficient way of attaining purely end-grafted polymer chains is to<br />
use thiol-terminated macromolecules, i.e. polymer chains ending in –SH,<br />
<strong>and</strong> gold substrates. In this case, one achieves a quite strong chemical bond<br />
(b)
230 8. ATOMIC FORCE MICROSCOPy ANd POLyMERS ON SURFACES<br />
(a)<br />
(b)<br />
FIgurE 8.3 Schematic drawing of a polymer brush in (a) good solvent conditions <strong>and</strong><br />
(b) poor solvent conditions. Upon the change of the solvent conditions, the polymer chains<br />
could self-organise in nanoscale aggregates of small groups of polymer chains instead of<br />
collapsing individually.<br />
between the thiol <strong>and</strong> gold, end-grafting the chain, while its backbone is<br />
not affected by the inert Au surface (negligible physisorption). A system<br />
based on several molecular weights of various thiol-terminated polymers<br />
chemisorbed on gold substrates from dilute solutions at different incubation<br />
times was used in order to investigate the different structural regimes<br />
of such end-grafted polymer monolayers <strong>and</strong> submonolayers [18–20]. The<br />
gold substrates were immersed in the polymer solutions for a designated<br />
duration <strong>and</strong> upon removal from the solution were rinsed exhaustively<br />
with fresh toluene <strong>and</strong> subsequently dried under a stream of argon. The<br />
AFM imaging was performed in water in contact mode at very low force<br />
set-point; capillary forces were avoided since both the sample <strong>and</strong> the<br />
cantilever/tip were immersed in water.<br />
In Figure 8.4, we show a three-dimensional AFM topography map<br />
of a gold substrate sample prepared by immersion in a toluene solution<br />
of thiol-terminated polystyrene, PS-SH (concentration 0.1 mg ml –1 ,<br />
weight-average molecular weight M w� 144 000 g mol –1 ), for 45 min. It<br />
was imaged by AFM in water (bad solvent for PS). The gold substrate<br />
morphology of flat gold terraces separated by valleys <strong>and</strong> channels can<br />
be observed, while well-separated globular polymer isl<strong>and</strong>s on top of the<br />
terraces can be clearly seen.<br />
Although the height of the polymeric isl<strong>and</strong>s can be deduced directly<br />
from AFM topography images, their lateral size is usually overesti-<br />
mated owing to the convolution between the volumes of the tip apex<br />
<strong>and</strong> the polymeric isl<strong>and</strong>. One can employ deconvolution methodologies
1000 nm<br />
8.3 ENd-gRAFTEd POLyMER CHAINS 231<br />
500 nm<br />
0 nm 0 nm<br />
500 nm<br />
24.95 nm<br />
1000 nm<br />
FIgurE 8.4 AFM 3D-graph topography image of a gold substrate immersed in a toluene<br />
solution of PS-SH (concentration: 0.1 mg ml –1 , molecular weight: 144 000 g mol –1 ) for 45 min.<br />
The AFM imaging was conducted in water, which is a bad solvent for PS. The polymer chains<br />
are in collapsed globular state <strong>and</strong> r<strong>and</strong>omly distributed on the gold terraces.<br />
including simple geometrical formulas in order to estimate the true volume<br />
of the isl<strong>and</strong>s [18, 20]. The dimensions of polymer isl<strong>and</strong>s deduced<br />
directly from AFM images show an average height <strong>and</strong> width of about<br />
5.5 <strong>and</strong> 48 nm, respectively. As we have explained, although the measurement<br />
of height can be reliable, the width is usually overestimated<br />
owing to the convolution effect. Assuming that the shape of a polymer<br />
isl<strong>and</strong> is a spherical cap <strong>and</strong> taking into account the convolution effect,<br />
the average volume of the globules was estimated to be around 310 nm 3 .<br />
This compares well with the volume of a single collapsed polymer chain,<br />
which can be calculated assuming the usual bulk density for polystyrene<br />
(�1 g cm –3 ). Therefore, we conclude that the polymer isl<strong>and</strong>s are isolated<br />
single polymer chains chemisorbed <strong>and</strong> collapsed into dense globules<br />
onto the gold substrate. This implies that for this incubation time the<br />
polymer chains are sparsely distributed <strong>and</strong> the grafting density is low<br />
enough for us to observe the ‘mushroom’ regime. The variation in the<br />
size of the polymer globules reflects the polymer polydispersity (M w/<br />
M n� 1.2, where M n is the number-average molecular weight) <strong>and</strong> possibly<br />
small changes in the globular conformation, which can be magnified<br />
by the convolution effect.<br />
Figures 8.5 <strong>and</strong> 8.6 show two examples of gold substrates that<br />
were immersed in a toluene solution of higher concentration of PS-SH<br />
(2 mg ml –1 ) for much longer incubation times (2 days) <strong>and</strong> imaged under<br />
water in contact AFM mode. We can clearly see that the gold substrates<br />
are covered by a dense array of large polymer globules. Forty-eight hours<br />
is adequate time for the adsorption to reach its maximum grafting density,<br />
<strong>and</strong> we expect to have several thous<strong>and</strong>s of chains within a square
232 8. ATOMIC FORCE MICROSCOPy ANd POLyMERS ON SURFACES<br />
1000 nm<br />
500 nm<br />
0 nm 0 nm<br />
500 nm<br />
34.56 nm<br />
1000 nm<br />
FIgurE 8.5 AFM 3D-graph topography image of a gold substrate immersed in a toluene<br />
solution of PS-SH (concentration: 2 mg ml –1 , molecular weight: 258 000 g mol –1 ) for 48 h.<br />
The AFM imaging was conducted in water, which is a bad solvent for PS. The polymer<br />
chains form aggregates (pinned micelles) on the gold terraces.<br />
1000 nm<br />
500 nm<br />
0 nm 0 nm<br />
500 nm<br />
29.25 nm<br />
1000 nm<br />
FIgurE 8.6 AFM 3D-graph topography image of a gold substrate immersed in a toluene<br />
solution of PS-SH (concentration: 2 mg ml –1 , molecular weight: 51 500 g mol –1 ) for 48 h.<br />
The AFM imaging was conducted in water, which is a bad solvent for PS. The polymer<br />
chains form aggregates (pinned micelles) on the gold terraces.<br />
micrometre area. The globular isl<strong>and</strong>s are too many <strong>and</strong> too large in size<br />
to be individual collapsed chains.<br />
Since for sufficiently long incubation time we expect to be close <strong>and</strong><br />
above the overlap grafting density, the polymer chains are too close with<br />
each other to collapse separately. Instead, they collapse fusing together in<br />
small clusters, the size of which is moderated by the grafted tethers. They<br />
form aggregates connected to the surface through the tethers; these structures<br />
have been called ‘pinned micelles’ or ‘octopus surface micelles’ <strong>and</strong><br />
they have been predicted by theory [21, 22].
1000 nm<br />
500 nm<br />
8.3 ENd-gRAFTEd POLyMER CHAINS 233<br />
0 nm<br />
0 nm 500 nm 1000 nm<br />
29 nm<br />
0 nm<br />
FIgurE 8.7 AFM topography image of a collapsed PS-SH brush that was formed in poor<br />
solvent conditions (concentration: 0.1 mg ml –1 , molecular weight: 144 000 g mol –1 , incubation<br />
time: 1 h). The imaging has been performed in contact mode under water (bad solvent for PS).<br />
The polymeric isl<strong>and</strong>s can become more extended if the grafting density<br />
is increased further. This can be attained by using a poorer solvent<br />
for the polymer solution; under poorer solvent conditions, the size of<br />
polymer chains is smaller <strong>and</strong> they can be chemisorbed much closer to<br />
each other. Figure 8.7 shows an AFM 3D-topography image of PS-SH<br />
chains end-grafted on the gold substrate from a cyclohexane (poor solvent<br />
for PS at room temperature) solution. In contrast with the well-separated<br />
polymer isl<strong>and</strong>s formed when the polymer monolayers were prepared in<br />
good solvent conditions, polymer monolayer formation under poor solvent<br />
conditions results in elongated <strong>and</strong> semi-continuous isl<strong>and</strong>s upon<br />
the evaporation of the solvent.<br />
The lateral inhomogeneity manifested in polymer nanostructures<br />
<strong>and</strong> nanopatterns <strong>and</strong> clearly observed in our experiments is consistent<br />
with theory <strong>and</strong> simulations for end-grafted polymer chains at relatively<br />
high <strong>and</strong> uniform grafting densities [23–28]. It is to be expected that<br />
even higher grafting densities will result in a homogeneous layer upon<br />
the change of solvent conditions or drying of the solvent. However, such<br />
grafting densities are difficult to attain by a ‘grafting to’ technique like the<br />
one we employed owing to the steric hindrance of preadsorbed chains.<br />
Polymer brushes of very high grafting densities can be attained by ‘grafting<br />
from’ synthetic techniques, where the polymer chains grow from<br />
low–molecular weight initiators that are immobilized on the surface at<br />
very high density [17].<br />
A polymer brush of uniform density in good solvent conditions forms<br />
a homogeneous layer that is stretched away from the surface owing to
234 8. ATOMIC FORCE MICROSCOPy ANd POLyMERS ON SURFACES<br />
the repulsive excluded volume interactions between the polymer segments.<br />
The stretching of the chains is moderated by their entropic elasticity.<br />
Owing to the thermal agitation <strong>and</strong> fluctuations, a polymer chain adopts<br />
an overall shape that allows it to maximise the number of possible conformations.<br />
Stretching decreases this number, which leads to an entropy<br />
loss, <strong>and</strong> for this reason a restoring force of entropic origin develops that<br />
resists deformation. A polymer brush in good solvent conditions should<br />
behave like a compliant/soft (non-linear) spring resisting compression,<br />
protecting the surface <strong>and</strong> even exhibiting low adhesive <strong>and</strong> frictional<br />
forces. An atomic force microscope allows us to probe directly the relevant<br />
nanomechanical properties at the local level.<br />
We prepared a polymer brush system by immersing gold substrates to<br />
a dilute aqueous thiol-terminated poly(methacrylic acid), PMAA-SH, solution<br />
for 24 h (M w� 66 200 g mol –1 , M w/M n� 2). We used AFM force spectroscopy<br />
(forward <strong>and</strong> reverse force distance curves) in deionised water (good<br />
solvent for PMAA) <strong>and</strong> acquired many force distance profiles. Two types of<br />
force profiles were observed on these monolayers. Figure 8.8(a) shows an<br />
example of the first type. During the forward movement of the tip towards<br />
the surface (approach), we observe the gradual increasing of the loading<br />
force owing to the non-linear compressive elasticity of the brush. No adhesive<br />
force is observed because the brush steric repulsion force screens the<br />
interaction of the tip with the substrate (in the same way that polymer<br />
chains anchored on colloidal particles stabilise their suspensions). During<br />
the reverse movement, there is no adhesion <strong>and</strong> the polymer brush is<br />
decompressed in a largely reversible way. Figure 8.8(b) shows that a<br />
simple exponential does describe the AFM data in the regime of intermediate<br />
compression of the brush.<br />
The Alex<strong>and</strong>er–de Gennes theory of a polymer brush [29–31] predicts<br />
a certain relationship for its compression that can be approximated by<br />
the following exponential for intermediate compressions (0.2�D/L 0�0.9)<br />
F (nN)<br />
(a)<br />
1.5<br />
1.0<br />
0.5<br />
0.0<br />
0 20 40 60<br />
D (nm)<br />
Forward<br />
Reverse<br />
80 100<br />
F (nN)<br />
(b)<br />
10 0<br />
10 –1<br />
10 –2<br />
0 20 40 60 80 100<br />
D (nm)<br />
FIgurE 8.8 (a) First type of force profile observed over a polymer brush. Notice the<br />
absence of hysteresis. (b) Logarithmic plot of the forward force distance curve.
8.3 ENd-gRAFTEd POLyMER CHAINS 235<br />
with a spherical probe: F(D) � exp(�2�D/L 0), where F(D) is the force<br />
exerted on the probe by the brush at a separation D between the probe<br />
surface <strong>and</strong> the solid substrate <strong>and</strong> L 0 is the brush thickness.<br />
We can directly check this equation on our data. We fitted several force<br />
profiles with the above equation <strong>and</strong> we obtained a brush thickness in the<br />
range of 100 nm, which is of the same order (albeit a bit larger) of the distance<br />
D, where we observe in force profiles the initial increase of the force,<br />
signifying the extent (thickness) of the brush away from the surface. The<br />
discrepancy could arise from the fact that since our probe is small, a large<br />
proportion of the chains lie near the edge <strong>and</strong> therefore are expected to<br />
bend than to compress. A more appropriate theory must take into account<br />
the spherical shape of the tip <strong>and</strong> the lateral bending of the chains.<br />
The second type of force profile is shown in Figure 8.9(a). The force<br />
profile during probe-tip approach is essentially the same: gradual increase<br />
of the loading force due to gradual compression of the brush. However,<br />
during the reverse force profile as the tip moves away from the surface,<br />
we observe some long-range attractive forces. The full length of one chain<br />
should be around 120 nm (calculated using the number-average molecular<br />
weight M n), which is consistent with the stretching events occurring up<br />
to approximately 200 nm. The magnitude of the attractive forces is in the<br />
range of 0.1–0.5 nN. Forces in this range can be attributed to physisorption<br />
of several monomers to a surface. Since we observe a gradual increase in<br />
the first derivative of these forces, we attribute them on parts of chains,<br />
individual chains <strong>and</strong>/or small clusters of chains that were adhered<br />
non-specifically to the tip <strong>and</strong> subsequently stretched during the reverse<br />
movement. The continuous lines in Figure 8.9 correspond to the entropic<br />
force for a single polymer chain using the freely jointed chain (FJC) model<br />
with full length approximately equal to the distance between the contact<br />
F (nN)<br />
1.5<br />
1.0<br />
0.5<br />
0.0<br />
Forward<br />
Reverse<br />
Theory<br />
–0.5<br />
50<br />
0 50 100<br />
D (nm)<br />
150<br />
(a) (b)<br />
D (nm)<br />
80<br />
70<br />
60<br />
Non-linear elasticity<br />
0.00 0.05 0.10 0.15<br />
F (nN)<br />
FIgurE 8.9 (a) Second type of force curve observed over a polymer brush. The continuous<br />
lines correspond to the theoretical prediction for stretching of individual chains or<br />
parts of a chain. (b) Zoom of one of the long-range attractive force curves.
236 8. ATOMIC FORCE MICROSCOPy ANd POLyMERS ON SURFACES<br />
point (D � 0) <strong>and</strong> the minimum of each attractive force. We have assumed<br />
that the Kuhn statistical segment is b � 0.6 nm. In Figure 8.9(b), we zoom<br />
on an attractive peak (dashed rectangular in Figure 8.9a). For this plot, the<br />
variables have been reversed (y � D <strong>and</strong> x � F) <strong>and</strong> the absolute value<br />
of the force has been used. The theoretical formula of the stretching by<br />
⎡<br />
its ends of a freely jointed chain is<br />
⎛ Fb⎞<br />
kT ⎤<br />
R � Lc<br />
⋅ ⎢coth<br />
⎜ �<br />
⎝⎜<br />
kT ⎠⎟<br />
⎥ , where R is<br />
⎣⎢<br />
Fb ⎦⎥<br />
the end-to-end chain distance when a restoring force F is exerted, Lc is<br />
the contour length (full length of the stretched polymer chain), k is the<br />
Boltzmann constant <strong>and</strong> T is the absolute temperature [32]. As we can see<br />
in Figure 8.8, it fits the AFM data well.<br />
8.4 DIBLoCk CoPoLymErS ADSorBED on SurFACES<br />
Amphiphilic diblock copolymers have great potential in a variety of<br />
applications, ranging from their use as responsive layers for the fabrication<br />
of smart surfaces to supramolecular responsive carriers for drug/<br />
gene delivery. We have recently studied the adsorption of poly(isopreneb-ethylene<br />
oxide) block copolymer (PI-PEO) on mica [33]. The polymer<br />
consists of a short <strong>and</strong> very flexible hydrophobic block (PI) (in melt state<br />
at room temperature) <strong>and</strong> a long hydrophilic block (PEO). Tapping mode<br />
AFM imaging was employed in dry state.<br />
The deposition of a droplet of deionised water polymer solutions<br />
onto freshly cleaved mica, followed by gentle rinsing <strong>and</strong> finally<br />
drying, resulted in the initial formation of ultraflat polymeric isl<strong>and</strong>s<br />
(Figure 8.10a). It is well known that mica is hydrophilic <strong>and</strong> an adsorbed<br />
water layer in the form of semi-continuous water isl<strong>and</strong>s forms on its<br />
fresh surface upon cleavage under normal ambient conditions [34, 35].<br />
The polymer molecules organised in polymer brush-like flat nanometre-thick<br />
isl<strong>and</strong>s with the PEO within the ultrathin water layer <strong>and</strong> the<br />
PI block collapsed at the water/air interface (Figure 8.11a). As mica<br />
became gradually less hydrophilic with time, most of the water evaporated<br />
<strong>and</strong> the polymeric monolayer isl<strong>and</strong>s were laterally compressed<br />
<strong>and</strong> increased in height (Figure 8.10b). Ultimately, parts of the PEO<br />
blocks remained adsorbed on to mica (some strongly adsorbed water<br />
molecules still remain on the mica surface) <strong>and</strong> other parts aggregated<br />
owing to the monomer-monomer attractive interactions in the dry state<br />
<strong>and</strong> dewetted the mica surface. The dewetting behaviour was aided further<br />
by the aggregation <strong>and</strong> fusion of the flexible PI blocks; they came<br />
together to create a hydrophobic core. The height of the polymer isl<strong>and</strong>s<br />
was increased by a factor of about 5 as mica became less hydrophilic with<br />
time. The final structure has the organisation of a surface micelle with
1000.0<br />
950.0<br />
900.0<br />
850.0<br />
800.0<br />
750.0<br />
700.0<br />
650.0<br />
Y [nm]<br />
–2300.0 –2200.0 –2100.0 –2000.0 –1900.0 –1800.0<br />
X [nm]<br />
8.5 STAR-SHAPEd POLyMERS AdSORBEd ON SURFACES 237<br />
5.0<br />
2.0<br />
0.0<br />
(a) (b)<br />
Y [nm]<br />
(c)<br />
2600.0<br />
2500.0<br />
Z [nm]<br />
Y [nm]<br />
600.0<br />
500.0<br />
400.0<br />
a hydrophobic core at the top <strong>and</strong> a stretched hydrophilic layer partially<br />
in close contact with the mica surface (Figures 8.10c <strong>and</strong> 8.11b).<br />
8.5 StAr-ShAPED PoLymErS ADSorBED on SurFACES<br />
Star-shaped polymers consist of several linear polymer chains (arms)<br />
covalently attached to a low–molecular weight core. They are of particular<br />
interest in soft condensed matter since the number of arms (also called<br />
functionality) controls their interactions <strong>and</strong> consequently many physical<br />
properties. It is expected that as the number of arms increases, their polymer-<br />
like behaviour will change <strong>and</strong> become more colloidal-like.<br />
We adsorbed polybutadiene (PB) star-shaped polymers of three different<br />
functionalities (number of arms: 18, 32 <strong>and</strong> 59) onto freshly cleaved<br />
mica surfaces from dilute polymer solutions in toluene. We investigated<br />
the structural regimes of the submonolayers <strong>and</strong> monolayers of adsorbed<br />
PB star polymers, which were formed on the mica surfaces by using AFM<br />
tapping mode imaging in dry state [36]. We attained various adsorbed<br />
300.0<br />
–2000.0 –1900.0 –1800.0 –1700.0 –1600.0 –1500.0 –1400.0<br />
X [nm]<br />
2400.0<br />
15.0<br />
12.5<br />
2300.0<br />
10.0<br />
7.5<br />
2200.0<br />
5.0<br />
2.5<br />
0.0<br />
–1800.0 –1700.0 –1600.0 –1500.0 –1400.0 –1300.0 –1200.0<br />
X [nm]<br />
17.5<br />
20.0<br />
Z [nm]<br />
10.0<br />
7.5<br />
5.5<br />
2.5<br />
0.0<br />
FIgurE 8.10 AFM 3D-graph topography of a polymer isl<strong>and</strong> (a) 133 min, (b) 745 min<br />
<strong>and</strong> (c) 2980 min after sample preparation.<br />
Z [nm]
238 8. ATOMIC FORCE MICROSCOPy ANd POLyMERS ON SURFACES<br />
Pl<br />
PEO<br />
(a)<br />
(b)<br />
PEO<br />
Pl<br />
FIgurE 8.11 Schematic drawing of the transition over time of the (a) initial brush-like<br />
flat polymer isl<strong>and</strong> to (b) a surface micelle. In (a), the PEO blocks are swollen <strong>and</strong> stretched<br />
away due to the presence of a water layer on mica surface <strong>and</strong> the resulting repulsive<br />
excluded volume interactions. In (b), most of water has evaporated <strong>and</strong> parts of the PEO<br />
blocks remain adsorbed on to the mica surface while other parts come together due to monomer-monomer<br />
attractions. The PI blocks fuse together in a compact core.<br />
amounts by employing several incubation times of the mica surface<br />
within the solution.<br />
Figure 8.12 shows three examples of topographies of samples prepared<br />
using relatively short incubation times <strong>and</strong> consequently low adsorbed<br />
amounts. In this case, the polymer isl<strong>and</strong>s correspond to single <strong>and</strong> isolated<br />
polymer chains, as can be deduced by an analysis of their size from<br />
the AFM images. We can clearly see the asymmetric/irregular shape of<br />
the low-functionality star polymers <strong>and</strong> the gradual transition to more<br />
circular shapes as the functionality increased. Furthermore, the height of<br />
the individual star polymers increased with functionality; notice the flat<br />
‘pancake’-like conformation taken by the low-functionality star polymers<br />
<strong>and</strong> the increased height of the high-functionality ones. It is clear that the<br />
higher the functionality, the closer is the star polymer behaviour to the<br />
one expected for a (soft) colloidal particle; their shape was affected much<br />
less strongly by the interactions with the surface.<br />
Figure 8.13 shows an example of an image of sample prepared using<br />
incubation times in the solution long enough for the 18-arm star polymers<br />
to start interacting as they were swollen in good solvent conditions.<br />
Owing to the increased adsorbed amount, i.e. increased number of chains<br />
adsorbed on the surface, they compressed <strong>and</strong> confined each other,<br />
decreasing their lateral extension <strong>and</strong> simultaneously increasing their<br />
height. In this way, because of this confinement effect, they became uniformly<br />
spaced, more symmetrical <strong>and</strong> circular in shape. At these moderate
Y Range: 1.21 μm<br />
(a)<br />
0.205<br />
–1.00 –0.399<br />
8.5 STAR-SHAPEd POLyMERS AdSORBEd ON SURFACES 239<br />
Height range: 2.234 nm<br />
1.03 1.63 2.24<br />
X Range: 1.21 μm<br />
Y Range: 1.21 μm<br />
(c)<br />
–1.71 –1.11<br />
–2.32<br />
2<br />
1.6<br />
1.2<br />
0.8<br />
0.4<br />
Height range: 5.050 nm<br />
–1.26 –0.652 –0.0473<br />
X Range: 1.21 μm<br />
adsorbed amounts, no interpenetration <strong>and</strong> fusion were observed. The star<br />
polymers continued to keep their individuality <strong>and</strong> they formed isolated<br />
globules upon the evaporation of the solvent.<br />
In Figure 8.14, we compare the overall monolayer organisation for<br />
long incubation times. Such incubation times are long enough to achieve<br />
the maximum adsorbed amount (within our adsorption protocol using<br />
dilute polymer solutions). The 18-arm star polymers organised in discontinuous<br />
asymmetric isl<strong>and</strong>s, which consist of several individual polymer<br />
chains; it is clear that the star polymers penetrated each other when in<br />
good solvent <strong>and</strong> they collapsed in groups upon drying. The 32-arm star<br />
polymers also penetrated each other <strong>and</strong> they self-assembled in a continuous<br />
dense network. On the other h<strong>and</strong>, the 59-arm star polymers did<br />
Y Range: 1.21 μm<br />
(b)<br />
–0.56 0.0452 0.65<br />
Height range: 10.03 nm<br />
–2.59 –1.98 –1.38<br />
X Range: 1.21 μm<br />
FIgurE 8.12 AFM topography images of star polymers adsorbed on mica: (a) 18 arms,<br />
incubation time of 5 min; (b) 32 arms, incubation time of 10 min <strong>and</strong> (c) 59 arms, incubation<br />
time of 180 min.<br />
10<br />
7.5<br />
5<br />
2.5<br />
5<br />
4<br />
3<br />
2<br />
1
240 8. ATOMIC FORCE MICROSCOPy ANd POLyMERS ON SURFACES<br />
Y Range: 1.21 μm<br />
–0.313<br />
–0.918<br />
–1.52<br />
Height range: 7.591 nm<br />
0.634 1.24 1.84<br />
X Range: 1.21 μm<br />
FIgurE 8.13 AFM topography image of an 18-arm star polymer adsorbed on mica,<br />
incubation time of 20 min.<br />
not interpenetrate. The large number of arms <strong>and</strong> the resulting crowding<br />
effect <strong>and</strong> increased osmotic pressure within their volume did not permit<br />
interpenetration between the star polymers <strong>and</strong> they continued to collapse<br />
individually despite the high adsorbed amount. This is a further<br />
manifestation of colloidal behaviour.<br />
8.6 ConCLuSIonS<br />
We have provided several examples of the use of AFM in investigations<br />
of the fine structure <strong>and</strong> local nanomechanical properties of polymer<br />
monolayers <strong>and</strong> submonolayers. AFM is ideally suited to study the lateral<br />
inhomogeneities of such ultrathin coatings <strong>and</strong> their non-linear compliance<br />
<strong>and</strong> adhesion either in dry state or within liquid environment.<br />
Although the time resolution of the (classical) AFM imaging modes (tapping<br />
mode in particular) is not particularly high, the molecular kinetics<br />
of polymers on surfaces can be sluggish enough to allow the real-time/<br />
real-space monitoring of various physico-chemical surface processes. As<br />
miniaturisation of electronic <strong>and</strong> medical devices is rapidly approaching<br />
the nanometre scale, the AFM is gradually becoming the most important<br />
characterisation tool of nanostructural <strong>and</strong> nanomechanical properties.<br />
Beyond the unique characterisation capabilities, systematic AFM studies<br />
on model systems can ultimately lead to a formulation of design criteria<br />
6<br />
4<br />
2
Y Range: 1.21 µm<br />
(a)<br />
0.403<br />
–0.201<br />
–0.806<br />
Height range: 8.035 nm<br />
–0.67 –0.0649 0.54<br />
X Range: 1.21 µm<br />
Y Range: 1.21 µm<br />
(c)<br />
–1.14 –0.536 –0.0684<br />
–0.461<br />
ACkNOwLEdgEMENTS 241<br />
8<br />
6<br />
4<br />
2<br />
Height range: 12.45 nm<br />
0.144 0.749<br />
X Range: 1.21 µm<br />
for the processing of surfaces with polymers at the nanoscale for various<br />
engineering applications.<br />
ACknowLEDgEmEntS<br />
The author would like to thank all who have contributed to the<br />
research reviewed in this chapter, in particular Georges Hadziioannou<br />
(who also introduced me to this fascinating subject matter), Eric van der<br />
Vegte, Emmanouil Glynos, Stergios Pispas, Alex<strong>and</strong>ros Chremos, Philip<br />
Camp, George Petekidis <strong>and</strong> Jacques Roovers.<br />
Y Range: 1.22 µm<br />
(b)<br />
–0.976<br />
–1.59<br />
–2.20<br />
Z range: 4.352 nm<br />
0.341 0.953 1.56<br />
X Range: 1.22 µm<br />
FIgurE 8.14 AFM topography images of star polymers adsorbed on mica: (a) 18 arms,<br />
incubation time of 2880 min; (b) 32 arms, incubation time of 2440 min <strong>and</strong> (c) 59 arms, incubation<br />
time of 3600 min.<br />
10<br />
7.5<br />
5<br />
2.5<br />
4<br />
3<br />
2<br />
1
242 8. ATOMIC FORCE MICROSCOPy ANd POLyMERS ON SURFACES<br />
LISt oF ABBrEvIAtIonS<br />
AFM atomic force microscope<br />
FJC freely jointed chain<br />
PB polybutadiene<br />
PEO poly(ethylene oxide)<br />
PI polyisoprene<br />
PI-PEO poly(isoprene-b-ethylene oxide)<br />
PMAA poly(methacrylic acid)<br />
PMAA-SH thiol-terminated poly(methacrylic acid)<br />
PS polystyrene<br />
PS-SH thiol-terminated polystyrene<br />
–SH thiol<br />
LISt oF SymBoLS<br />
Unit<br />
b Kuhn statistical segment M<br />
D separation distance between the probe surface<br />
<strong>and</strong> the solid substrate surface<br />
M<br />
F force N<br />
k Boltzmann’s constant (1.38 � 10�23 ) J K�1 Lc contour length of a polymer chain (full length<br />
of a stretched polymer chain)<br />
M<br />
L0 polymer brush thickness M<br />
Mn number-average molecular weight g mol –1<br />
Mw weight-average molecular weight g mol –1<br />
R end-to-end polymer chain distance M<br />
T absolute temperature K<br />
X x-coordinate or abscissa<br />
Y y-coordinate or ordinate<br />
references<br />
[1] J.I.I. Laco, F.C. Villota, F.L. Mestres, Corrosion protection of carbon steel with thermoplastic<br />
coatings <strong>and</strong> alkyd resins containing polyaniline as conductive polymer, Prog.<br />
Org. Coat. 52 (2005) 151–160.
REFERENCES 243<br />
[2] J.J. Licari, L.A. Hughes, H<strong>and</strong>book of Polymer Coatings for Electronics: Chemistry,<br />
Technology <strong>and</strong> Applications Materials Science <strong>and</strong> Process Technology Series,<br />
second ed., Noyes Publications, New York, 1990.<br />
[3] J. Lahann, Vapor-based polymer coatings for potential biomedical applications,<br />
Polym. Int. 55 (2006) 1361–1370.<br />
[4] M.H. Jung, H. Lee, Patterning of conducting polymers using charged self-assembled<br />
monolayers, Langmuir 24 (2008) 9825–9831.<br />
[5] M. Sjöberg, L. Bergström, A. Larsson, E. Sjöström, The effect of polymer <strong>and</strong> surfactant<br />
adsorption on the colloidal stability <strong>and</strong> rheology of kaolin dispersions, Colloids Surf.<br />
A Physicochem. Eng. Asp. 159 (1999) 197–208.<br />
[6] Q.Q. Li, M. Zaiser, V. Koutsos, Carbon nanotube/epoxy resin composites using a<br />
block copolymer as a dispersing agent, Physica Status Solidi A Appl. Res. 201 (2004)<br />
R89–R91.<br />
[7] M.J. Yang, V. Koutsos, M. Zaiser, Interactions between polymers <strong>and</strong> carbon nanotubes:<br />
a molecular dynamics study, J. Phys. Chem. B 109 (2005) 10009–10014.<br />
[8] Z. Zhang, H. Vaisocherová, G. Cheng, W. Yang, H. Xue, S.Y. Jiang, Nonfouling<br />
behavior of polycarboxybetaine-grafted surfaces: structural <strong>and</strong> environmental effects,<br />
Biomacromolecules 9 (2008) 2686–2692.<br />
[9] D. Pavithra, M. Doble, Biofilm formation, bacterial adhesion <strong>and</strong> host response on<br />
polymeric implants – issues <strong>and</strong> prevention, Biomed. Mater. 3 (2008) 034003.<br />
[10] A. Kikuchi, T. Okano, Intelligent thermoresponsive polymeric stationary phases<br />
for aqueous chromatography of biological compounds, Prog. Polym. Sci. 27 (2002)<br />
1165–1193.<br />
[11] T. Rohr, D.F. Ogletree, F. Svec, J.M.J. Fréchet, Surface functionalization of thermoplastic<br />
polymers for the fabrication of microfluidic devices by photoinitiated grafting, Adv. Funct.<br />
Mater. 13 (2003) 264–270.<br />
[12] L. Léger, E. Raphaël, H. Hervet, Surface-anchored polymer chains: their role in adhesion<br />
<strong>and</strong> friction, in: K. Binder <strong>and</strong> S. Granick (eds) Polymers in Confined Environments,<br />
Springer-Verlag, Berlin, 1999, pp. 185–225.<br />
[13] S.H. Anastasiadis, H. Retsos, S. Pispas, N. Hadjichristidis, S. Neophytides, Smart<br />
polymer surfaces, Macromolecules 36 (2003) 1994–1999.<br />
[14] C. Xu, X. Fu, M. Fryd, S. Xu, B.B. Wayl<strong>and</strong>, K.I. Winey, R.J. Composto, Reversible stimuli-<br />
responsive nanostructures assembled from amphiphilic block copolymers, Nano Lett.<br />
6 (2006) 282–287.<br />
[15] P.M. Mendes, Stimuli-responsive surfaces for bio-applications, Chem. Soc. Rev. 37<br />
(2008) 2512–2529.<br />
[16] G. Binnig, C.F. Quate, C. Gerber, Atomic Force Microscope, Phys. Rev. Lett. 56 (1986)<br />
930–933.<br />
[17] B. Zhao, W.J. Brittain, Polymer brushes: surface-immobilized macromolecules, Prog.<br />
Polym. Sci. 25 (2000) 677–710.<br />
[18] V. Koutsos, E.W. van der Vegte, P.C.M. Grim, G. Hadziioannou, Isolated polymer<br />
chains via mixed self-assembled monolayers: morphology <strong>and</strong> friction studied by<br />
scanning force microscopy, Macromolecules 31 (1998) 116–123.<br />
[19] V. Koutsos, E.W. van der Vegte, G. Hadziioannou, Direct view of structural regimes of<br />
end-grafted polymer monolayers: a scanning force microscopy study, Macromolecules<br />
32 (1999) 1233–1236.<br />
[20] V. Koutsos, E.W. van der Vegte, E. Pelletier, A. Stamouli, G. Hadziioannou, Structure<br />
of chemically end-grafted polymer chains studied by scanning force microscopy in<br />
bad-solvent conditions, Macromolecules 30 (1997) 4719–4726.<br />
[21] D.R.M. Williams, Grafted polymers in bad solvents – octopus surface micelles, Journal<br />
de Physique II 3 (1993) 1313–1318.<br />
[22] E.B. Zhulina, T.M. Birshtein, V.A. Priamitsyn, L.I. Klushin, Inhomogeneous structure of<br />
collapsed polymer brushes under deformation, Macromolecules 28 (1995) 8612–8620.
244 8. ATOMIC FORCE MICROSCOPy ANd POLyMERS ON SURFACES<br />
[23] P.Y. Lai, K. Binder, Structure <strong>and</strong> dynamics of polymer brushes near the theta point –<br />
a Monte-Carlo Simulation, J. Chem. Phys. 97 (1992) 586–595.<br />
[24] G.S. Grest, M. Murat, Structure of grafted polymeric brushes in solvents of varying<br />
quality – a molecular-dynamics study, Macromolecules 26 (1993) 3108–3117.<br />
[25] H. Tang, I. Szleifer, Phase-behavior of grafted polymers in poor solvents, Europhys.<br />
Lett. 28 (1994) 19–24.<br />
[26] C. Yeung, A.C. Balazs, D. Jasnow, Lateral Instabilities in a grafted layer in a poor<br />
solvent, Macromolecules 26 (1993) 1914–1921.<br />
[27] K.G. Soga, H. Guo, M.J. Zuckermann, Polymer brushes in a poor solvent, Europhys.<br />
Lett. 29 (1995) 531–536.<br />
[28] R.S. Ross, P. Pincus, Bundles – end-grafted polymer layers in poor solvent, Europhys.<br />
Lett. 19 (1992) 79–84.<br />
[29] S. Alex<strong>and</strong>er, Adsorption of chain molecules with a polar head – a scaling description,<br />
Journal de Physique 38 (1977) 983–987.<br />
[30] P.G. de Gennes, Conformations of polymers attached to an interface, Macromolecules<br />
13 (1980) 1069–1075.<br />
[31] P.G. de Gennes, Polymers at an interface – a simplified view, Adv. Colloid Interface<br />
Sci. 27 (1987) 189–209.<br />
[32] A.Y. Grosberg, A.R. Khokhlov, Statistical physics of macromolecules, AIP Press,<br />
Woodbury, NY, 1994.<br />
[33] E. Glynos, S. Pispas, V. Koutsos, Amphiphilic diblock copolymers on mica: formation of<br />
flat polymer nanoisl<strong>and</strong>s <strong>and</strong> evolution to protruding surface micelles, Macromolecules<br />
41 (2008) 4313–4320.<br />
[34] J. Hu, X.-d. Xiao, D.F. Ogletree, M. Salmeron, Imaging the condensation <strong>and</strong> evaporation<br />
of molecularly thin films of water with nanometer resolution, Science 268 (1995)<br />
267–269.<br />
[35] J. Hu, X.-d. Xiao, D.F. Ogletree, M. Salmeron, Structure of molecularly thin films of<br />
water on mica in humid environments, Surf. Sci. 344 (1995) 221–236.<br />
[36] E. Glynos, A. Chremos, G. Petekidis, P.J. Camp, V. Koutsos, Polymer-like to soft colloidlike<br />
behavior of regular star polymers adsorbed on surfaces, Macromolecules 40 (2007)<br />
6947–6958.
C H A P T E R<br />
9<br />
Application of Atomic<br />
Force Microscopy for<br />
the Study of Tensile <strong>and</strong><br />
Microrheological Properties<br />
of Fluids<br />
Matthew S. Barrow <strong>and</strong> P. Rhodri Williams<br />
O U T L I N E<br />
9.1 Introduction 245<br />
9.2 Dynamic AFM Methods for the Characterisation of<br />
Material Properties 249<br />
9.3 Dynamic Modulation Studies on Confined Fluids 252<br />
9.4 Determination of Rheological Properties from Resonance Spectra 256<br />
9.5 Cavitation <strong>and</strong> Adhesive Failure of Thin Films 259<br />
9.6 Mesoscale Experimental Studies of the Tensile Behaviour<br />
of Thin Fluid Films 261<br />
Acknowledgements 269<br />
List of Symbols 269<br />
References 270<br />
9.1 InTRoDuCTIon<br />
Atomic force microscopy is undeniably one of the most successful techniques<br />
for the characterisation of surfaces <strong>and</strong> is routinely used to describe<br />
structural details with nanoscale resolution [1]. The ability of atomic force<br />
245
246 9. APPLICATION OF ATOMIC FORCE MICROSCOPy<br />
microscopy to differentiate between local mechanical properties is well<br />
known, e.g. phase contrast maps obtained using tapping mode [2, 3] may<br />
be used to identify the distribution of components in composite materials.<br />
Alternatively, ‘adhesive’ forces may be determined by monitoring the<br />
deflection of a calibrated cantilever during contact <strong>and</strong> withdrawal of a<br />
probe from the surface of a ‘tacky’ material. During this process, the contact<br />
time <strong>and</strong> applied force may be varied, <strong>and</strong> by performing multiple<br />
tests at different locations, a representative ‘adhesive force map’ may be<br />
constructed. The atomic force microscope (AFM) is clearly a powerful tool<br />
for the investigation of forces which govern the mechanics of processes<br />
occurring at or below the microscale, <strong>and</strong> the exceptional ability of atomic<br />
force microscopy to determine forces associated with the microscale deformation<br />
<strong>and</strong> flow of fluids is presently discussed.<br />
An underst<strong>and</strong>ing of the rheology of complex fluids is of fundamental<br />
importance in many practical engineering <strong>and</strong> biomedical applications.<br />
Traditional rheometrical techniques, e.g. cone <strong>and</strong> plate rheometers, require<br />
reasonably large volumes (i.e. several millilitres) of test fluid, which is often<br />
undesirable as limited volumes of fluid may be available, e.g. in studies of<br />
biological fluids. Many processes also involve the deformation of fluids in<br />
geometric confinement, which are not clearly described by the macroscale<br />
or bulk rheology alone, e.g. thin film lubrication, <strong>and</strong> in such instances, it<br />
is desirable to characterise the fluid mechanics under ostensibly similar<br />
conditions. Additionally, in such situations in situ measurement of fluid<br />
rheology is impaired by either the confined space or the extreme process<br />
environments.<br />
The rheological behaviour of thin liquid films is an important aspect<br />
of lubrication [4] <strong>and</strong> printing [5] – processes which often involve mesoscale<br />
(0.1�10 �m) thickness films undergoing rapid deformation between<br />
separating surfaces. In the case of fluid mechanical machinery, they are<br />
usually solid surfaces whereas in biomechanics they may be flexible<br />
surfaces such as biological membranes. The ‘cracking’ of knuckle joints has<br />
been attributed to cavitation within mesoscale lubricating films of synovial<br />
fluid [6] whereas surface damage in microelectromechanical systems<br />
devices has been attributed to the cavitation of lubricant films [7]. The<br />
ability of liquids to sustain tension is an important factor in the survival<br />
of plants in which the cohesion–tension (C–T) theory has been proposed<br />
to explain water transport [8]. The C–T theory assumes that water, when<br />
confined in small tubes with wettable walls such as xylem elements, can<br />
sustain a tension ranging from 3 to 30 MPa. The liquid forms a continuous<br />
system in the water-saturated cell walls from the evaporating surfaces of<br />
the leaves to the absorbing surfaces of the roots. During evaporation, the<br />
reduction in water potential at the surfaces causes movement of water out<br />
of the xylem, with water loss producing tension in the xylem sap that is<br />
transmitted throughout the continuous water columns to the roots.
9.1 INTROdUCTION 247<br />
In coating processes, cavitational film splitting may result in the formation<br />
of rapidly stretching filaments, whose breakup leads to unwanted<br />
droplet deposition. Filament formation is also a feature of coating flows<br />
involving adhesive films [9], but descriptions of the process are still<br />
largely qualitative, invoking terms such as ‘tack’ [10, 11]. The tack of an<br />
ink film is primarily connected with the tensile forces developed in film<br />
splitting [12], the function of cavitation being to limit the forces of separation<br />
of surfaces joined by a tacky liquid [13]. By definition, the tack of<br />
an ink film is the maximum tensile stress (or ‘negative pressure’) which<br />
can be withstood by it before splitting [5].<br />
The subject area of micro- <strong>and</strong> nanorheology [14–18] has developed<br />
rapidly <strong>and</strong> benefited from the advent of the AFM <strong>and</strong> the surface force<br />
apparatus (SFA) [19, 20]. Roughly speaking, there are two principle objectives<br />
driving the development of AFM-based rheometrical techniques, the<br />
first considers the ability of AFM to describe the microrheological properties<br />
of thin fluid films [21] whereas the second seeks to exploit AFM<br />
microcantilever technology as the basis for microsensor devices [22],<br />
which, e.g. may be incorporated into the so-called lab on a chip device.<br />
Specifically, it is envisaged that the development of MEMS devices such<br />
as diagnostic sensors incorporating microfluidic components can greatly<br />
benefit from an improved underst<strong>and</strong>ing of the microscale rheology.<br />
In this respect, the determination of viscosity <strong>and</strong> density (or even simple<br />
representative damping coefficients) inferred via microflexural responses is<br />
of fundamental interest. This is because microcantilevers present a viable<br />
means by which the rheological properties of microvolumes of fluids may be<br />
assessed in situ. As an example, the resistive forces arising due to the motion<br />
of a cantilever within a minute volume of liquid can be analysed through<br />
changes in the cantilever response, as an immersed cantilever undergoing<br />
stimulated oscillation will experience fluid damping effects which are interrogable<br />
via resonance characteristics [23, 24]. Similarly, the ability to detect<br />
adsorbed mass (by monitoring changes in the resonance characteristics) in<br />
conjunction with functionalisation of the cantilever (or probe) surface represents<br />
a potential approach to biomolecule or chemical species detection [25].<br />
As the forces accompanying indentation or compression of a material<br />
by an AFM probe may be resolved with extreme accuracy, AFM-based<br />
techniques are particularly suited to rheological studies of predominantly<br />
elastic materials [26, 27]. In the simplest case, cantilever deflection may<br />
be used to describe the compliance of the surface in a simple qualitative<br />
manner, e.g. for a rigid, essentially undeformable material, the deflection<br />
of the cantilever will be equivalent to the downwards displacement of the<br />
piezoceramic actuator whereas for softer materials the probe will compress<br />
<strong>and</strong> indent the sample such that the cantilever deflection is lower.<br />
Probe–surface interactions assuming ideal elastic interactions are often<br />
used to determine the Young’s modulus of the material via derivatives of
248 9. APPLICATION OF ATOMIC FORCE MICROSCOPy<br />
the Hertz model (i.e. sphere against sphere compression) for non-adhesive<br />
systems. Furthermore, the Hertz model is readily adapted to provide<br />
idealised descriptions of geometries such as conical tips compressing flat<br />
substrates [28].<br />
Notwithst<strong>and</strong>ing the fact that AFM is routinely used in studies where<br />
elastic deformations are satisfactorily described by idealised models – a<br />
limitation of the use of AFM in many rheological experiments, particularly<br />
those involving fluids, is the assumed description of both the geometry<br />
<strong>and</strong> the flow field contained within.<br />
Therefore, for the AFM-based study of fluids, the use of colloid probes<br />
[29] has proved to be of considerable benefit. Colloid probes (which are<br />
fabricated by attaching a colloid sphere to the underside of an AFM cantilever)<br />
have found favour in several different research areas including the<br />
drainage of thin films [30], interfacial forces [31], mechanical properties<br />
of cells [32–34], lubrication theory [35] <strong>and</strong> the deformation of colloidal<br />
droplets [36, 37] to name but a few.<br />
As has been alluded to previously, the successful use of AFM in<br />
microrheological <strong>and</strong> tribological studies is facilitated by the ability to<br />
describe force through the deflection <strong>and</strong> known spring constant of the<br />
cantilever. However, the topographical imaging capability of an AFM has<br />
been used as a component part of some studies. The imaging capabilities<br />
of an AFM have been used to measure the biaxial extensional deformation<br />
involved in the industrially important process of bubble inflation [38].<br />
In this method the force-measuring capabilities of the AFM are not fully<br />
exploited, instead the AFM is used to measure the radius of curvature of<br />
microbubbles produced by inflating a thin polymer film, which is adhered<br />
to a porous silicon substrate. A similar approach is used to study the surface<br />
properties of polymers by using an AFM to measure the embedding<br />
rate of nanoparticles at temperatures near to the glass transition temperature<br />
of the polymer [39].<br />
Considering force measurement, one technique of particular relevance<br />
to tribology <strong>and</strong> lubrication studies is lateral force measurements in friction<br />
force microscopy (FFM) [40–44], which considers the torsion of a cantilever<br />
arising from frictional forces generated between a scanning tip <strong>and</strong> a surface.<br />
As lateral perturbations are caused by both frictional forces <strong>and</strong> surface<br />
features, it is informative to consider the local surface topography. The influence<br />
of surface features upon the lateral deflection becomes apparent if the<br />
surface is scanned in opposing directions, as a topography-induced torsional<br />
response is dependent upon the scan direction (the torsion is reversed)<br />
whereas the lateral deflection arising from the effect of the frictional component<br />
will remain unchanged (Figure 9.1).<br />
The beneficial application of microrheometry as an adjunct to traditional<br />
bulk rheometry is demonstrated by the application of AFM force<br />
studies (including dynamic FFM, [45]) to adhesive materials. For example,
Cantilever<br />
Scan direction<br />
9.2 dyNAMIC AFM METHOdS 249<br />
Friction force<br />
FIguRE 9.1 FFM monitors the lateral deflection, i.e. tilt or ‘twist’ of the cantilever due<br />
to frictional resistance or drag forces in thin fluid layers acting upon the scanning tip.<br />
a combination of lateral force microscopy <strong>and</strong> indentation has been used<br />
to investigate the surface properties of pressure-sensitive adhesives (PSAs)<br />
[46], where the tack of the adhesive in response to various compressive<br />
loads is of paramount importance. Differences in local frictional forces due<br />
to increased adhesive force between the tip <strong>and</strong> the PSA arising from the<br />
incorporation of tack promoting components (tackifiers) may be identified.<br />
Therefore, FFM may be used to first identify microscale regions of varying<br />
surface friction, following which the AFM can determine ‘adhesive’<br />
characteristics of the different tackifier enriched or depleted regions. The<br />
correlation between traditional bulk tack <strong>and</strong> adhesion force as measured<br />
by AFM (or nanotack) is also considered by other workers [47], who find<br />
that the AFM force data are in general agreement with bulk tack tests.<br />
9.2 DynAMIC AFM METhoDS FoR ThE<br />
ChARACTERISATIon oF MATERIAL PRoPERTIES<br />
The nanorheological properties of polymeric liquids can be obtained<br />
by adapting techniques such as an SFA or an AFM to act in a dynamic<br />
mode [48, 49]. However, the results of AFM studies are often more<br />
difficult to interpret than those derived from SFA experiments due to<br />
uncertainties about the zero separation distance, the influence of probe<br />
asperities <strong>and</strong> torsional deflections.<br />
Dynamic AFM, especially force modulation microscopy (FMM) [50–55],<br />
as a technique to determine the dominant elastic (<strong>and</strong> also viscoelastic)<br />
properties of materials has been investigated in many studies including<br />
the drainage of confined fluid layers [56], elastohydrodynamic lubrication<br />
studies [57], approximations to the viscoelastic moduli of solvated<br />
polymer layers <strong>and</strong> the evaluation of temperature-induced variations in<br />
micromechanical properties [58]. Force modulation studies involve applying<br />
a small amplitude (typically a few nanometres), sinusoidal oscillation<br />
Tilt
250 9. APPLICATION OF ATOMIC FORCE MICROSCOPy<br />
to either the probe or the sample substrate <strong>and</strong> observing the amplitude<br />
<strong>and</strong> phase response of the cantilever when the probe is either in contact<br />
with the surface of the material or is in close proximity to a surface<br />
with an intervening fluid layer. Although many forms of the modulation<br />
technique have been employed to characterise predominantly elastic<br />
materials, comparatively few AFM studies use superimposed modulation<br />
to study confined liquids.<br />
Various terms are used to describe dynamic modes of operation<br />
although many of the reported modes are essentially similar in design.<br />
Tapping mode (or intermittent contact) imaging is perhaps the most familiar<br />
dynamic AFM technique. In tapping mode the cantilever is oscillated<br />
at or near its resonance frequency such that when the tip interacts with the<br />
sample surface, the amplitude <strong>and</strong> phase response will change. In response<br />
to this deviation, a feedback loop adjusts the height of cantilever above the<br />
surface to achieve a constant vibration amplitude. Phase contrast images<br />
obtained from tapping mode studies are routinely discussed in terms of<br />
the perceived damping effects induced by the elasticity or compliance of<br />
the sample surface. Most commonly, qualitative interpretation of the phase<br />
data may be used to discern areas of varying stiffness. FMM is essentially<br />
a dynamic mode of operation in which the tip <strong>and</strong> substrate interact under<br />
conditions of a constant (average) force such that the amplitude of the<br />
cantilever oscillation varies about a nominal set point. Modulated force<br />
experiments are often performed at a single point upon the surface, <strong>and</strong><br />
for the case of intervening liquids, the amplitude <strong>and</strong> phase are often monitored<br />
as a function of probe to surface distance.<br />
Considering predominantly elastic surfaces, FMM is often used to<br />
generate surface images, which are generally referred to as either ‘elasticity<br />
maps’ or ‘viscoelasticity maps’, depending upon the interpretation<br />
of the force data. The fundamental approach employs single-point force<br />
modulation [59] in which an ‘AC modulation’ is superimposed upon<br />
either the probe or the sample during a ‘DC’ (i.e. force–distance) experiment.<br />
If the modulation frequency is sufficiently high, the technique may<br />
be employed while scanning a surface <strong>and</strong> is a complementary approach<br />
to point-wise force mapping [28].<br />
The basic method <strong>and</strong> interpretation is considered by Maivald et al.<br />
(1991) [52] who exploit variations in local surface elasticity to provide a<br />
basis for image contrast in inhomogeneous materials. The probe is scanned<br />
across the surface with the feedback loop maintaining contact with a<br />
(nominal) constant cantilever deflection <strong>and</strong> therefore a constant applied<br />
force. If the sample is caused to oscillate a small distance �z 1 in the direction<br />
normal to the surface, the undulating motion of the surface produces<br />
a corresponding superimposed deflection of the cantilever �z 2. For an<br />
infinitely hard sample, the deflection of the cantilever will be such that<br />
�z 1 � �z 2, whereas in the case of softer samples the sample surface will
9.2 dyNAMIC AFM METHOdS 251<br />
be compressed by a distance �z 3 such that �z 1 � �z 2 � �z 3. Therefore<br />
the deflection of the cantilever �z 2 is less on softer materials, <strong>and</strong> the term<br />
�z 2/�z 1 is deemed to be indicative of surface compliance (Figure 9.2).<br />
This method can be implemented using the advanced capabilities<br />
of modern AFMs, one approach is similar to that described by Scott <strong>and</strong><br />
Bushan (2003) [2]. In this example, the modulating amplitude is obtained<br />
by inducing vibration of the cantilever, the sample remaining fixed in position.<br />
Using an interleaved scanning approach, the sample height is first<br />
determined using tapping mode. The probe height is then adjusted such<br />
that (i) the probe is kept in constant contact with the surface <strong>and</strong> (ii) a constant<br />
scan height is maintained by compensating for the known sample<br />
height variation, i.e. the distance between the fixed end of the cantilever<br />
<strong>and</strong> the surface is held constant. The surface is then rescanned with the<br />
oscillating probe driven, in this case, at the resonant frequency (Figure 9.3).<br />
Clearly, the method by which the sample–tip interaction is spatially<br />
modulated can be achieved in several ways. The probe may be caused<br />
1<br />
1 2<br />
FIguRE 9.2 FMM: discrimination between areas of varying stiffness. (1) Hard substrate<br />
<strong>and</strong> (2) soft substrate; cantilever deflection signal �z 2 is reduced due to deformation of the<br />
sample.<br />
2
252 9. APPLICATION OF ATOMIC FORCE MICROSCOPy<br />
Surface profile from primary scan<br />
FIguRE 9.3 Interleaved scanning wherein the force modulation scan uses the predetermined<br />
sample height.<br />
to oscillate by (i) the influence of an external driving force (such as a<br />
magnetic field), (ii) vibrating the sample using an electromechanical<br />
oscillator, (iii) mounting the cantilever on a secondary piezoceramic oscillator<br />
or (iv) modifying the primary piezoceramic scanner voltage. The<br />
preferred method may be influenced by the apparatus <strong>and</strong> the desired<br />
frequency range, as some oscillators exhibit mechanical resonances <strong>and</strong><br />
produce excessive acoustic interference, the latter problem is likely to be<br />
more pronounced when the sample is contained in a small liquid cell.<br />
A wide range of modulation frequencies <strong>and</strong> experimental methods have<br />
been employed, ranging from a few hertz to several megahertz, <strong>and</strong><br />
although the basic excitation principle is retained, the analytical methods<br />
vary significantly. For example, atomic force acoustic microscopy<br />
(AFAM) employs excitation frequencies to several megahertz in order to<br />
elicit sample-specific, frequency-dependent spectra derived from contact<br />
measurements [60, 61].<br />
9.3 DynAMIC MoDuLATIon STuDIES on<br />
ConFInED FLuIDS<br />
For those studies where the viscoelastic response is of interest,<br />
attempts are made to relate the deflection amplitude <strong>and</strong> phase of the<br />
cantilever to the dynamic complex modulus G*(�) of the material. In<br />
st<strong>and</strong>ard rheometry, a sample subjected to a sinusoidally varying shear<br />
stress will respond with a sinusoidally varying shear strain, <strong>and</strong> the<br />
mechanical characteristics of the sample may then be described by the<br />
complex modulus as given by<br />
�( �)<br />
G*( �)<br />
�<br />
ε( �)<br />
(9.1)
where � is the shear stress, ε the shear strain <strong>and</strong> � the frequency. The<br />
complex modulus may be described as the summation of an in-phase<br />
elastic component <strong>and</strong> an out-of-phase viscous component, i.e.<br />
G*( �) � G′ ( �) � i G′′<br />
( �)<br />
(9.2)<br />
<strong>and</strong><br />
9.3 dyNAMIC MOdULATION STUdIES ON CONFINEd FLUIdS 253<br />
G′<br />
( �)<br />
� tan �<br />
G′′<br />
( �)<br />
(9.3)<br />
where G� is the storage (elastic) modulus, G�″ the loss (viscous) modulus<br />
<strong>and</strong> � the phase angle between stress <strong>and</strong> strain.<br />
The analogy between bulk oscillatory <strong>and</strong> dynamic scanning probe<br />
microscope techniques is discussed by Suraya et al. (2008) [62], where<br />
for the latter, the motion of one surface imparts oscillatory motion to<br />
the fluid due to hydrodynamic forces, which in turn invokes oscillation of<br />
the ‘receiver’ surface, i.e. the colloid probe. However, although the receiver<br />
surface will oscillate at the same frequency, viscous effects are such that the<br />
oscillation amplitude is attenuated <strong>and</strong> the phase is shifted (relative to the<br />
driven surface). For an inelastic liquid, the detected signal will be 90° outof-phase,<br />
whereas an elastic response is expected to be in-phase with the<br />
driven surface. When oscillating surfaces bearing adsorbed polymer layers<br />
are immersed in dilute polymer solutions, the vibrating motion produces<br />
a viscous response at large surface separations. As the separation distance<br />
is reduced, the polymer layers interact <strong>and</strong> deform, <strong>and</strong> the viscoelastic<br />
properties of the material become apparent as the receiver amplitude<br />
increases <strong>and</strong> the phase shift reduces.<br />
Suraya et al. (2008) note that the measurement of the effect of the hydrodynamic<br />
force upon the phase <strong>and</strong> amplitude is insufficient to directly<br />
measure the stress <strong>and</strong> strain in the fluid layer. To translate the measured<br />
parameters into terms which are physically representative of the stress<br />
<strong>and</strong> strain, a mechanical model is required. A common approach employs<br />
hydrodynamic lubrication equations, which conveniently describe the<br />
hydrodynamic force that arises from the flow of a purely viscous liquid<br />
between a flat surface <strong>and</strong> a colloid sphere, the force F is given by<br />
F U R<br />
� 6 π η , ( h
254 9. APPLICATION OF ATOMIC FORCE MICROSCOPy<br />
The application of hydrodynamic lubrication approximations (equation<br />
(9.4)) is illustrated by the viscous damping coefficient described in<br />
the work of Friedenberg <strong>and</strong> Mate (1996) [63] who studied the amplitude<br />
<strong>and</strong> phase response of a colloid probe in contact with a thin film of lowmolecular-weight<br />
poly(dimethylsiloxane) (PDMS). The PDMS film was<br />
constrained between a sphere attached to a tungsten cantilever <strong>and</strong> a flat<br />
substrate – the substrate being subjected to oscillatory motion. Both the<br />
separation distance between the sphere <strong>and</strong> the surface h <strong>and</strong> the vibration<br />
frequency � were varied (Figure 9.4).<br />
A simple viscous model incorporating the contribution of meniscus<br />
forces is considered, which relates both amplitude <strong>and</strong> phase to a single,<br />
viscous damping coefficient b p. Equations (9.5) <strong>and</strong> (9.6) show the simplest<br />
form of the derived model when capillary forces are neglected.<br />
� φ<br />
� � bp<br />
k<br />
sin<br />
tanφ<br />
b<br />
p<br />
�<br />
�<br />
k<br />
bp 2<br />
6π�R<br />
�<br />
h<br />
(9.5)<br />
(9.6)<br />
(9.7)<br />
where � is the ratio of cantilever <strong>and</strong> drive amplitudes, k <strong>and</strong> φ the spring<br />
constant <strong>and</strong> the phase difference between the substrate <strong>and</strong> cantilever<br />
motion <strong>and</strong> R the radius of the sphere.<br />
PDMS fluid<br />
Silicon<br />
substrate<br />
Tungsten<br />
cantilever tip<br />
FIguRE 9.4 Schematic diagram of the probe–fluid geometry (Friedenberg <strong>and</strong> Mate,<br />
1996, the illustrated geometry has been rotated 90°). As the probe is not fully immersed,<br />
capillary forces affect the contact area.<br />
h<br />
dz
The value of the viscous damping coefficient b p is then determined by<br />
fitting the measured values of both amplitude <strong>and</strong> phase as a function of<br />
separation distance h to equations (9.5) <strong>and</strong> (9.6). The apparent viscosity<br />
of the fluid � may then be calculated from the damping coefficient <strong>and</strong><br />
equation (9.7). Using this method reasonable agreement between the calculated<br />
viscosity <strong>and</strong> measured ‘bulk’ viscosity was found; however, the<br />
results also indicated that the surface roughness of the sphere may cause<br />
deviations from the expected zero-intercept relationship inferred from<br />
equation (9.7) (i.e. 1/b p vs. h). Choi <strong>and</strong> Kato (2003) [64] have extended<br />
this approach to investigate the shear properties of perfluoropolyether<br />
liquid bridges formed between two glass spheres by inducing lateral oscillations<br />
(as opposed to the normal perturbations depicted in Figure 9.4).<br />
Suraya et al. (2008) [62] have reported the use of dynamic AFM to<br />
investigate the viscoelastic response of adsorbed hydroxypropyl guar<br />
(HPG) layers. The results identified dominant viscous properties at large<br />
surface separations <strong>and</strong> viscoelastic behaviour at closer separations<br />
where the polymer layers interact. The adsorbed polymer also exhibited<br />
viscoelastic frequency dependence <strong>and</strong> ‘shear thinning’ characteristics.<br />
Similar studies were performed by Braithwaite <strong>and</strong> Luckham (1999)<br />
[65] who studied the viscoelastic response of adsorbed layers of gelatine<br />
under compression using dynamic modulation. The system consisted<br />
of a colloid probe interacting with a hard substrate where both surfaces<br />
were coated with a thin film of adsorbed gelatine. The model employed<br />
considers the characteristics of the viscoelastic materials as discussed by<br />
Radmacher et al. (1993) [66].<br />
In such studies, as the actual values of stress <strong>and</strong> strain in the sample<br />
cannot be determined directly, they are instead speculatively related to<br />
the force F <strong>and</strong> displacement z of the cantilever through the use of an<br />
‘apparatus coefficient’ b such that for a given frequency:<br />
�<br />
ε<br />
� � G<br />
1 F<br />
*<br />
b z<br />
(9.8)<br />
From which an alternative definition of the storage <strong>and</strong> loss moduli may<br />
be described [65, 66] wherein the equations are of the form:<br />
<strong>and</strong><br />
9.3 dyNAMIC MOdULATION STUdIES ON CONFINEd FLUIdS 255<br />
�k (cos φ � �)<br />
G′<br />
( �)<br />
�<br />
2 b � � 2 � cosφ<br />
� 1<br />
�k sin φ<br />
G′′<br />
( �)<br />
�<br />
2 b � � 2 � cosφ<br />
� 1<br />
(9.9)<br />
(9.10)
256 9. APPLICATION OF ATOMIC FORCE MICROSCOPy<br />
where k is the cantilever force constant, � the ratio of cantilever to substrate<br />
modulation amplitude <strong>and</strong> φ the phase angle between substrate<br />
<strong>and</strong> cantilever response. It is important to note that the equivalence of<br />
the phase terms φ <strong>and</strong> � is not always physically representative, <strong>and</strong> a<br />
quantitative description of the rheological properties of a homogeneous<br />
fluid sample may be verified under some conditions; however, when the<br />
sample fluid is non-uniform, e.g. an adsorbed polymer layer is present,<br />
only qualitative information can be obtained.<br />
9.4 DETERMInATIon oF RhEoLogICAL PRoPERTIES<br />
FRoM RESonAnCE SPECTRA<br />
The resonance characteristics of microcantilevers immersed in fluids<br />
have long been recognised as a potential basis for rheometrical microsensor<br />
technology as the resonance spectra of a vibrating cantilever are<br />
influenced by the viscosity <strong>and</strong> density of the surrounding medium.<br />
Compared to vibration in air, the fundamental resonant frequency of a<br />
rectangular cantilever is reduced when operated in liquid; immersion<br />
in increasingly more viscous liquids reduces the resonant frequency<br />
further – this is accompanied by a broadening of the resonant peak <strong>and</strong> a<br />
decrease in peak amplitude (Figure 9.5).<br />
The shift in the mechanical response is attributed to the viscous<br />
drag <strong>and</strong> inertial effects caused by the fluid adjacent to the cantilever. The<br />
resonance characteristics are often derived from the effect of environmental<br />
stimuli – most commonly thermally induced vibrations. The thermal<br />
noise spectrum is derived from the Fourier transform of the displacement<br />
of the cantilever. Fast Fourier transforms (FFTs) are routinely used<br />
in sound <strong>and</strong> vibration analysis to convert a signal from a time domain to<br />
A<br />
High viscosity<br />
ω<br />
Low viscosity<br />
FIguRE 9.5 Resonance frequency shift due to immersion in viscous fluids.
a frequency domain such that the relative magnitude of each frequency<br />
component may be evaluated <strong>and</strong> the resonance spectrum constructed.<br />
The motion of the cantilever may be modelled as a simple harmonic<br />
oscillator (SHO) with an increased effective mass due to the contribution<br />
of the fluid in close proximity to the beam. Simple approximations<br />
describing the contribution of the fluid mass are described by Oden et al.<br />
(1996) [83], who model the virtual mass of the SHO as the combination<br />
of the cantilever mass <strong>and</strong> the mass of a spherical fluid volume enveloping<br />
the cantilever. More advanced models are considered by Sader (1998)<br />
[67], who present analytical expressions for the hydrodynamic functions,<br />
which determine the hydrodynamic loading due to the motion of the<br />
mass of fluid around a rectangular beam. The theory considers cantilevers<br />
with rectangular <strong>and</strong> cylindrical cross-sections for which the length<br />
is far greater than the diameter or width, respectively.<br />
For beams of a circular cross-section, the analytical expression for the<br />
hydrodynamic function � has been established previously; however, rectangular<br />
beams of finite thickness are far more complicated to describe.<br />
Sader (1998) derives a correction factor �(�) which relates the circular<br />
cross-section solution to that of an infinitely thin rectangular cantilever<br />
based on the approximate relationship previously described by Tuck<br />
(1969) [69] such that the hydrodynamic function � rect is given by:<br />
� ( �) � �( �) � ( �)<br />
rect circ<br />
where the correction factor �(�) is of the form<br />
<strong>and</strong><br />
9.4 dETERMINATION OF RHEOLOgICAL PROPERTIES FROM RESONANCE SPECTRA 257<br />
�( �) � � ( �) � i�<br />
( �)<br />
r i<br />
� ( �), � ( �) � (Re)<br />
r i f<br />
(9.11)<br />
(9.12)<br />
(9.13)<br />
For small amplitude oscillations, the characteristic Reynolds number<br />
depends upon the cantilever width x such that:<br />
� �<br />
Re �<br />
4�<br />
fluid x2<br />
For an SHO the amplitude A(�) is given by<br />
A0�R<br />
A(<br />
�)<br />
≅<br />
⎡<br />
⎢ 2 2 � �<br />
⎢(<br />
� �R)<br />
⎣<br />
Q<br />
2<br />
� � 2<br />
2<br />
1<br />
2 2 2<br />
R<br />
⎤<br />
⎥<br />
⎦<br />
(9.14)<br />
(9.15)
258 9. APPLICATION OF ATOMIC FORCE MICROSCOPy<br />
where A 0 is the zero-frequency amplitude response, � <strong>and</strong> � R are the<br />
radial <strong>and</strong> radial resonant frequencies, respectively [70],<br />
<strong>and</strong><br />
�<br />
R<br />
�<br />
�<br />
⎡<br />
⎢<br />
π�x<br />
⎢<br />
1 �<br />
⎣ 4�<br />
vacuum<br />
2<br />
⎤<br />
� r( �R)<br />
⎥<br />
⎦<br />
4� � �r(<br />
�<br />
2 R)<br />
π�x<br />
Q �<br />
� ( � )<br />
i R<br />
1 2<br />
(9.16)<br />
(9.17)<br />
where � (�� cxt) is the mass per unit length of the cantilever of density � c,<br />
width x <strong>and</strong> thickness t. � r <strong>and</strong> � i are the real <strong>and</strong> imaginary parts of the<br />
aforementioned hydrodynamic function �(�).<br />
For an unknown fluid, the fundamental resonance profile is used to<br />
determine � R <strong>and</strong> Q from equation (9.15). The method is not restricted to<br />
the use of the fundamental resonance mode, <strong>and</strong> higher resonance peaks<br />
may be used although the accuracy of the result is reduced. Provided<br />
� vacuum is known, the experimentally determined values of the resonance<br />
frequency <strong>and</strong> the quality factor may then be used to determine the viscosity<br />
<strong>and</strong> density by numerically solving equations (9.16) <strong>and</strong> (9.17).<br />
The SHO model is useful for the determination of both viscosity <strong>and</strong><br />
density provided Q � 1. If the cantilever response is heavily damped,<br />
Q � 1, then the SHO model is not valid. However, this may in some<br />
instances be circumvented by the selection of a cantilever with different<br />
mechanical properties.<br />
Maali et al. (2005) [71] have evaluated a simplified approximation for<br />
the hydrodynamic function �(�) <strong>and</strong> propose that<br />
<strong>and</strong><br />
�r( �)<br />
� a � a<br />
1 2<br />
�<br />
�<br />
� ⎛�<br />
⎞<br />
�i( �)<br />
� a3 � a ⎜ 4<br />
� ⎝⎜<br />
�⎠⎟<br />
2<br />
(9.18)<br />
(9.19)<br />
where a 1, a 2, a 3 <strong>and</strong> a 4 are constants <strong>and</strong> � is the characteristic thickness of<br />
fluid surrounding the cantilever equivalent to an exponential damping<br />
length �, where<br />
� �<br />
2�<br />
� �<br />
fluid
9.5 CAvITATION ANd AdHESIvE FAILURE OF THIN FILMS 259<br />
The accuracy of the model proposed by Sader (1998) was evaluated<br />
using several fluids including air, water, acetone, carbontetrachloride<br />
<strong>and</strong> butanol, which presented wide viscosity <strong>and</strong> density ranges. Results<br />
were presented for both precision fabricated <strong>and</strong> st<strong>and</strong>ard cantilevers,<br />
<strong>and</strong> in both cases the theoretical <strong>and</strong> experimental results for the Q factor<br />
<strong>and</strong> resonant frequency as determined from the thermal resonance<br />
spectra were in close agreement [72] as were the viscosity <strong>and</strong> density<br />
results [70], which were found to be in accordance with known bulk<br />
measurements.<br />
The resonance methods described provide a basis for the in situ determination<br />
of both viscosity <strong>and</strong> density of inelastic liquids. However, the<br />
use of resonance data to elucidate viscoelastic parameters is extremely<br />
complicated. In this respect, the use of a modified Langevin model which<br />
incorporates a complex drag coefficient in an attempt to overcome the<br />
limitations of the simplified SHO model has also been studied [73–75].<br />
9.5 CAvITATIon AnD ADhESIvE FAILuRE<br />
oF ThIn FILMS<br />
Due to the high deformation rates which typify many mesoscale phenomena,<br />
such as film splitting, filamentation <strong>and</strong> cavitation processes,<br />
significant viscoelastic effects may be anticipated. One such effect is a<br />
delay in the cavitation of viscoelastic liquids in micrometre-sized gaps,<br />
due to the development of normal stresses [76]. Others claimed that<br />
viscoelastic effects include a displacement of the point of cavitation from<br />
the centre of contact (where film thickness is a minimum) <strong>and</strong> enhanced<br />
film thicknesses [77]. Little is known about the influence of viscoelasticity<br />
in sub-micron liquid film cavitation, but the initial film thickness is a<br />
crucial factor: for sufficiently thin films, even ostensibly low rates of<br />
surface separation may provoke the high rates of fluid deformation necessary<br />
to generate enough tension (through viscous forces) to result in<br />
cavitation [78].<br />
It is important to realise that in ultrathin films of water, cavitation may<br />
occur spontaneously, due to the antipathy between the liquid <strong>and</strong> hydrophobic<br />
surfaces between which it is confined [79]. Spontaneous cavitation<br />
was first observed experimentally by Christenson <strong>and</strong> Claesson (1988)<br />
[79]. Theory predicts that vaporous cavities will only form in pure liquids<br />
as a result of large tensions, some 1300–1400 bar in the case of water [80],<br />
although a somewhat higher figure (ca. 1900 bar) results from an interpretation<br />
of the thermodynamic properties of stretched water known<br />
as the stability limit conjecture [81]. Experiments involving very small<br />
quantities of pure water have produced tensions close to this homogeneous<br />
nucleation limit [82], but they are not commonly observed.
260 9. APPLICATION OF ATOMIC FORCE MICROSCOPy<br />
Berard et al. (1993) [83] showed that the free energy of a bridging cavity<br />
is lower than that of liquid water when the surfaces are separated as far as<br />
micrometres <strong>and</strong> claim that the fact that such cavities are not observed as<br />
the two surfaces approach contact from far apart indicates that the liquid<br />
between them is metastable, i.e. there is some barrier preventing cavitation.<br />
The theory for the long-ranged hydrophobic attraction relies upon<br />
this notion of induced cavitation – the force between two colloids in a<br />
near-critical or a near-spinodal fluid is attractive <strong>and</strong> long-ranged – <strong>and</strong><br />
the connection between the spinodal attractions in the bulk <strong>and</strong> measured<br />
long-range attractions between hydrophobic surfaces is the observed<br />
cavitation [84].<br />
Computer simulations by Berard et al. [83] on a Lennard–Jones liquid<br />
confined between hard walls showed cavitation at small separations, <strong>and</strong><br />
that there was indeed a spinodal separation. Approaching this separation<br />
it was found that the attractions were much stronger than the van der<br />
Waals attraction <strong>and</strong> longer ranged. Qualitatively, a separation-induced<br />
spinodal can account for the measured hydrophobic attractions.<br />
In the case of cavitation which results from the development of fluid<br />
mechanical stresses (as tension), Joseph [78] has pointed out that the concept<br />
of ‘negative pressure’ is not particularly useful; it is more pertinent to<br />
consider the state of stress experienced by a liquid. In doing so, it is convenient<br />
to decompose the stress into a deviator <strong>and</strong> mean normal stress,<br />
the most positive value of principal stresses being the maximum tension.<br />
In order to facilitate a comparison of a liquid’s cavitation threshold stress<br />
with the principal stresses at each point within the liquid, it is necessary<br />
to know the flow field. In terms of studying cavitation inception within<br />
mesoscale liquid films, this requirement imposes stringent experimental<br />
dem<strong>and</strong>s as it requires a comparison of the cavitation threshold at each<br />
point in a liquid sample with the principal stresses there. For liquids in<br />
motion, cavitation criteria must be based not on the pressure, but on the<br />
stress, <strong>and</strong> a cavitation bubble will open in the direction of maximum tension<br />
in principal coordinates. An important point which emerges is that<br />
a liquid can cavitate as a result of experiencing a shear deformation, the<br />
resulting cavity being pulled open by tension in the direction defined by<br />
principal stresses.<br />
Of the few experimental techniques capable of working at (or below)<br />
the mesoscale, the various ‘force microscopes’, such as the SFA, have been<br />
the most successful, but instances of their use in studies of thin fluid films<br />
are comparatively rare. Notable exceptions are provided by the work of<br />
Israelachvili <strong>and</strong> co-workers [85] who observed the growth <strong>and</strong> disappearance<br />
of vapour cavities in liquid films between separating mica surfaces in<br />
SFA experiments. The growth of a cavity was claimed to represent a ‘new’<br />
cavitation damage mechanism, insofar as surface damage occurred during<br />
cavity inception [86]. This is an interesting finding given that by far
9.6 MESOSCALE ExPERIMENTAL STUdIES 261<br />
the greatest effort in cavitation damage research has involved the study of<br />
bubble collapse <strong>and</strong> its consequences. Lord Rayleigh’s seminal analysis of<br />
the collapse of an isolated spherical void in an incompressible liquid [87]<br />
leads to the conclusion that, as the collapse nears completion, the pressure<br />
inside the liquid becomes indefinitely large. It is principally this ‘Rayleigh<br />
collapse’ mechanism (albeit extensively modified) which has led to the<br />
association of bubble collapse with cavitation damage [88].<br />
Little is known about the dynamics of cavities formed in thin films but,<br />
due to their inevitably close proximity to the film’s bounding surfaces, significant<br />
departures from spherical symmetry may be anticipated [89]. The<br />
asymmetry of cavity collapse leads to potentially damaging phenomena,<br />
such as liquid jets [90], but the issue of cavitation damage due to cavity<br />
growth has received relatively little attention, despite the possibly damaging<br />
consequences to capillaries <strong>and</strong> small blood vessels due to the intraluminal<br />
expansion of cavitation bubbles in the areas of laser angioplasty [91],<br />
electrohydraulic lithotripsy [92] <strong>and</strong> shock wave lithotripsy [93].<br />
9.6 MESoSCALE ExPERIMEnTAL STuDIES oF<br />
ThE TEnSILE BEhAvIouR oF ThIn FLuID FILMS<br />
As discussed, many processes of emerging scientific <strong>and</strong> technological<br />
interest involve the rheological behaviour of complex fluids in the mesoscale<br />
domain (ca. 0.1–10 �m), <strong>and</strong> in order to study the mesoscopic behaviour of<br />
fluids, particularly filament formation (i.e. extensional flow) <strong>and</strong> cavitationinduced<br />
failure <strong>and</strong> tack, it is necessary to satisfy some basic requirements.<br />
Notwithst<strong>and</strong>ing the fact that the creation of a liquid bridge between a colloid<br />
probe <strong>and</strong> a flat surface is a fairly straightforward process, manipulation<br />
of the desired quantity of fluid is not. Therefore in addition to recording<br />
the evolution of tensile forces, the deformation of the fluid should ideally be<br />
recorded. Although instruments such as the AFM suggest themselves for<br />
adaptation in terms of the former requirement, the latter task (i.e. recording<br />
the deformation of mesoscale filaments with sufficient temporal resolution)<br />
is non-trivial. But it is one which must be accomplished as a precursor to the<br />
development of a satisfactory mesoscale extensional flow technique.<br />
Extensional flows of complex fluids may be studied using several macroscale<br />
techniques such as the filament stretching rheometer (FSR) or the<br />
capillary break-up extensional rheometer (CaBER), whose illustrations are<br />
shown in Figures 9.6 <strong>and</strong> 9.7. In these techniques the characteristic dimensions<br />
<strong>and</strong> the quantity of liquid are known, whereas the interpretation of<br />
results from AFM-based microrheometry is restricted by the necessary<br />
assumptions about the flow field.<br />
As a colloid probe moves rapidly away from a surface, it is reasonable<br />
to assume that the fluid confined between the surface <strong>and</strong> the sphere may
262 9. APPLICATION OF ATOMIC FORCE MICROSCOPy<br />
FIguRE 9.6 Formation of macroscopic viscoelastic liquid filaments formed in a filament<br />
stretching rheometer. The fluid is constrained between the ends of two cylinders, which are<br />
pulled apart at a specified rate. The generated tensile force <strong>and</strong> the filament profile are used<br />
to calculate the extensional stress. The cylinder shown is 10 mm in diameter.<br />
FIguRE 9.7 Images showing a typical CaBER experiment on a biopolymer solution.<br />
The surfaces are separated at a predetermined distance, thereafter the temporal evolution of<br />
the surface profile is used to determine rheological properties. Cylinder diameter is 2 mm.<br />
be extended. If the initial minimum separation distance is sub-micron<br />
<strong>and</strong> the retraction event is sufficiently rapid, then the stresses generated<br />
within the liquid may approach or exceed the tensile strength of the liquid.<br />
However, in AFM studies the fluid is not directly observed, therefore<br />
fluid draining, viscoelastic filamentatious behaviour or cavitational failure<br />
is not readily substantiated. It is, therefore, useful to examine those<br />
means by which the deforming liquid can be observed <strong>and</strong> documented.<br />
The separation distances considered in most AFM force studies are near<br />
to, or beyond, the limits of conventional optical microscopy. Although it<br />
is not possible to observe nanofilaments through an optical microscope,
9.6 MESOSCALE ExPERIMENTAL STUdIES 263<br />
FIguRE 9.8 Photograph of unmodified (left) <strong>and</strong> modified (right) explorer AFMs.<br />
The base-section of the modified AFM has been modified to provide access for peripheral<br />
lenses.<br />
the observation of microfilaments is plausible. For cavitation <strong>and</strong> tackrelated<br />
phenomena, the rates of separation may be comparatively high,<br />
<strong>and</strong> in order to verify the existence of microfilaments <strong>and</strong> determine<br />
their form by direct visualisation, high-speed video microscopy studies<br />
are necessary.<br />
The combination of high-magnification microscopy <strong>and</strong> high-speed<br />
video presents a considerable challenge due to the illumination requirements,<br />
particularly when using high magnifications, optical filters (often<br />
necessary due to diffuse reflection of the AFM laser source) <strong>and</strong> fast<br />
shutter speeds. The design of most AFMs restricts direct observation of<br />
confined fluid samples, as the physical dimensions do not readily permit<br />
the inclusion of supplementary (high-speed) imaging systems. Several<br />
different AFMs have been used to study microfluidic aspects including<br />
a Thermomicroscopes Explorer, a Digital Instruments Dimension 3100,<br />
a Veeco Multimode <strong>and</strong> more recently a PSIA XE-120 with optical head.<br />
Studies using high-speed video systems, namely, a Kodak Ektapro 4540 mx<br />
<strong>and</strong> a Photron APX Fastcam, were used to record microfluidic defomation<br />
using ‘cold light’ sources to reduce the effect of evaporation when studying<br />
microlitres of aqueous polymer solutions.<br />
The enclosed nature of the Explorer prevents the use of peripheral<br />
lenses; therefore a section of the base of the AFM was removed (shown<br />
in Figure 9.8). For the Dimension 3100, the colloidal probe <strong>and</strong> cantilever<br />
are unobtrusively located beneath the scanner such that an objective lens
264 9. APPLICATION OF ATOMIC FORCE MICROSCOPy<br />
may be positioned adjacent to the cantilever holder. For the multimode<br />
system, a light intensified high-speed camera (an APX I2 intensified<br />
head) was employed due to illumination restrictions. Photomicrographic<br />
studies using the XE-120 (optical head system) are readily achieved as<br />
the optical head system permits peripheral access.<br />
The selection of a suitable microscope assembly is determined principally<br />
by the minimum attainable distance between lens <strong>and</strong> colloid<br />
probe; therefore ultralong working distance (ULWD) lenses were used<br />
(with supplementary lenses) as shown in Figure 9.9.<br />
Separation of the surfaces is achieved either by raising the probe (conventional<br />
force–distance) or by lowering the sample substrate using a<br />
piezoceramic actuator. The Explorer is especially suited to the latter as it<br />
may be positioned above an independent piezoceramic stack.<br />
Figure 9.10 shows filamentatious behaviour of a viscoelastic polymer<br />
(the V-shaped cantilever is 85 µm long). The fluid is stretched between<br />
FIguRE 9.9 The high-speed optical microscopy system incorporating (1) APX Fastcam<br />
high-speed camera, (2) extension tube, (3) supplementary macrolens magnification <strong>and</strong><br />
motorised focus, (4) light source <strong>and</strong> (5) ULWD microscope lens.<br />
FIguRE 9.10 Formation of microscale viscoelastic liquid filament formed between a<br />
V-shaped cantilever <strong>and</strong> a reflective silica substrate.
9.6 MESOSCALE ExPERIMENTAL STUdIES 265<br />
a ‘bare’ cantilever <strong>and</strong> a reflective silica substrate, the reflected image is<br />
apparent in the lower half of the figure. When probeless cantilevers are<br />
brought into contact with liquid layers or drops, the cantilever is often<br />
enveloped by the fluid. In this respect the use of a colloid probe can help<br />
prevent wetting of the cantilever beam <strong>and</strong> the associated degradation of<br />
the laser signal.<br />
Figure 9.11(a) shows a colloid probe attached to the tip of a V-shaped<br />
cantilever supported underneath the dimension scanner. In this example,<br />
no optical filters are used as careful alignment of both the optical microscope<br />
<strong>and</strong> the focal point of the laser reduces the transmission of diffuse<br />
laser light to the camera; although diffuse reflections are apparent (on the<br />
right h<strong>and</strong> side of the image), they are not excessive. If the observation<br />
angle is varied, such that the cantilever is viewed slightly from above,<br />
then either (i) optical filters are used or (ii) the field of view of the microscope<br />
is adjusted to omit all but the tip of the cantilever (under the same<br />
illumination conditions, omission of the optical filters permits higher<br />
photographic recording rates). This method is not employed using a light<br />
intesified system as the intensifier could be irreparably damaged. Figure<br />
9.11(b) shows the interaction between a 15-�m diameter colloid probe<br />
<strong>and</strong> a ‘large’ droplet of silicon oil, which is spread upon a reflective silica<br />
substrate. When first brought into contact, the silicon oil immediately<br />
enveloped the sphere <strong>and</strong> also adhered to the underside of the cantilever<br />
even though the cantilever was far above the surface of the drop; this<br />
effect was not readily apparent from above (as observed via the integrated<br />
AFM optics). The image shows the non-ideal liquid geometry during the<br />
(a) (b)<br />
FIguRE 9.11 (a) A colloid probe attached to the end of a V-shaped cantilever as<br />
observed using the optical system shown in Figure 9.9. The cantilever is observed directly<br />
from the side, hence the V-shape is not apparent. (b) The partial envelopment of a 15-�m<br />
diameter colloid probe by an excess of silicon oil (� � 12 Pa s).
266 9. APPLICATION OF ATOMIC FORCE MICROSCOPy<br />
late stages of separation of the surfaces, i.e. the liquid bridge is in contact<br />
with both the probe <strong>and</strong> the cantilever.<br />
Figure 9.12 shows the typical results obtained when attempting to generate<br />
a micro-CaBER experiment by the separation of a colloid probe <strong>and</strong><br />
a flat surface coated with PDMS silicon oil. In this case, the size of the<br />
FIguRE 9.12 Selected frames documenting the thinning of a liquid bridge formed by<br />
the microscale separation of surfaces.
9.6 MESOSCALE ExPERIMENTAL STUdIES 267<br />
sphere <strong>and</strong> the thickness of the fluid layer are such that only the underside<br />
of the colloid is wetted. In frame 01, meniscus effects can be seen,<br />
when the surfaces are separated, a microscale liquid bridge is formed,<br />
which persists at large separations. The liquid bridge thins until the filament<br />
fails between frames 07 <strong>and</strong> 08, leaving residual liquid on both the<br />
probe <strong>and</strong> the lower surface. The evolution of the filament profile closely<br />
resembles the macroscale CaBER results shown in Figure 9.7.<br />
Figure 9.13 shows the behaviour of a thin liquid film during the ‘highspeed’<br />
separation of the surfaces. In this instance, the high rates of separation<br />
<strong>and</strong> the high stress invoked in the fluid initially resist separation of the surfaces.<br />
Ultimately, the confined liquid layer appears to yield spontaneously<br />
(frame 02), forming a residual filament (initially 10-�m long <strong>and</strong> 1.3-�m in<br />
diameter created at an apparent rate � 5000 �m s�1 . The liquid filament then<br />
thins <strong>and</strong> breaks, residual liquid can be observed in frame 04.<br />
The adaptation of an AFM to act as a microrheometer has several<br />
potential benefits. In conventional rheometry, the generation of high<br />
rates of deformation is difficult, particularly so in extensional flow techniques<br />
where the defining strain rate is approximated by the relationship<br />
�ε � . U/h. The uniaxial rate of extension of cylindrical filaments between<br />
separating surfaces is equal to the ratio of the separation velocity U <strong>and</strong><br />
the instantaneous length of the filament h.<br />
FIguRE 9.13 The high-speed separation of surfaces bridged by a thin film of silicon oil<br />
(� � 12 Pa s). Frame interval is 2 m s.
268 9. APPLICATION OF ATOMIC FORCE MICROSCOPy<br />
Therefore when the initial surface separation distance h is large, i.e. the<br />
initial characteristic length of the fluid sample is �1 mm, the attainment<br />
of moderate extensional rates (�100 s �1 ) requires an initial separation<br />
velocity � 100 mm s �1 , which is readily achieved using linear drive systems.<br />
However, the filament length must increase exponentially as a function<br />
of time in order to maintain a constant rate of extension, a situation<br />
which most drive systems are unable to accommodate. Consequently,<br />
many tensile studies involving liquids are restricted to low rates of extension<br />
(�10 s �1 ) or small total strains. In this respect, the recreation of elongational<br />
flows at the mesoscale using piezoceramic technology presents<br />
one possible method by which high extensional strain rates <strong>and</strong> large total<br />
strains may be achieved. Furthermore, the difficulties encountered when<br />
attempting to measure rapidly changing, small tensile forces, which are<br />
generated within low viscosity, liquid filaments may also be addressed by<br />
utilising the precise force measuring capabilities of the AFM.<br />
The determination of tensile forces in flows which contain a dominant<br />
elongational component is readily achieved using an AFM force sensor;<br />
however, the ability to precisely determine elongational stress at large<br />
strains is only possible when the shape <strong>and</strong> size of the liquid layer is<br />
clearly understood. In this respect, it is necessary to document the evolution<br />
of a microfilament to ascertain the minimum diameter such that the<br />
maximum tensile stress may be calculated. This is particularly important<br />
as the transient evolution of the profile of Newtonian <strong>and</strong> non-Newtonian<br />
liquid bridges may vary significantly such that a representation of the<br />
tensile stress through terms such as F/R 2 may not be valid at large separations.<br />
This is because the critical dimension (i.e. the filament diameter) is<br />
not simply related to separation distance <strong>and</strong> is itself dependent upon the<br />
viscoelastic properties of the material.<br />
The ability of many fluid mechanical systems to perform the desired<br />
function is critically dependent upon the tensile properties of the material,<br />
<strong>and</strong> the inability of the fluid to sustain large tensions is often beneficial,<br />
e.g. in coating <strong>and</strong> lubricating flows. Characterisation of the shear, extensional<br />
<strong>and</strong> cavitational behaviour is clearly important, <strong>and</strong> the ability of<br />
an AFM to recreate high-rate deformations which are comparable to those<br />
encountered in industrial processes is of particular interest. Moreover, the<br />
ability to experimentally recreate film splitting, fibrillation <strong>and</strong> cavitational<br />
cohesive failure in one dynamic process is desirable. However, observation<br />
of rapid microscale phenomena is not straightforward, especially so<br />
when cavitational mechanisms are encountered, as the observation of the<br />
growth <strong>and</strong> collapse of vapourous cavities requires microsecond temporal<br />
resolution. Future work will extend the capabilities of the AFM–optical<br />
microscopy system to further investigate the presence <strong>and</strong> influence of cavitational<br />
effects prior to the apparent cohesive failure of lubricant layers. As<br />
such, it is necessary to evolve the design of the experiment <strong>and</strong> in the first
LIST OF SyMbOLS 269<br />
instance it is intended to conduct additional microscopic observations from<br />
beneath the liquid layer. In this way, the current restrictions imposed by the<br />
focal distance can be reduced (it is necessary to use ULWD dry lenses to<br />
document filament formation), <strong>and</strong> high numerical aperture oil immersion<br />
lenses may be employed to study microscale cavitational effects in the thin<br />
fluid layer prior to the mesoscale deformation process.<br />
ACknowLEDgEMEnTS<br />
The authors would like to acknowledge Prof. W. <strong>Richard</strong> <strong>Bowen</strong> <strong>and</strong><br />
Prof. <strong>Nidal</strong> <strong>Hilal</strong> for their significant contribution to the research content<br />
presented in section 9.6. The authors are grateful for the financial support<br />
of the EPSRC, the work reported here was conducted under EPSRC grant<br />
no. GR/S10438/01.<br />
LIST oF SyMBoLS<br />
A Cantilever vibration amplitude m<br />
an Constant<br />
B Geometric apparatus coefficient m<br />
bp Viscous damping coefficient N s m�1 F Force N<br />
G* Complex modulus Pa<br />
G’′ Elastic modulus Pa<br />
G“″ Viscous modulus Pa<br />
H Separation distance between probe <strong>and</strong> surface m<br />
K Spring constant N m�1 R Radius of sphere m<br />
T Cantilever thickness m<br />
U Relative separation velocity m s�1 X Cantilever width m<br />
Z Displacement m<br />
� Ratio of cantilever <strong>and</strong> drive amplitude<br />
�circ Hydrodynamic function for cylindrical cantilever<br />
�rect Hydrodynamic function for rectangular<br />
cantilever<br />
� Rheological phase angle °
270 9. APPLICATION OF ATOMIC FORCE MICROSCOPy<br />
ε Shear strain<br />
� Shear viscosity Pa s<br />
� Exponential damping length m<br />
� Mass per unit length of cantilever kg m �1<br />
� c Cantilever density kg m �3<br />
� fluid Fluid density kg m �3<br />
� Shear stress N m �2<br />
φ Phase difference between substrate <strong>and</strong> cantilever °<br />
� Frequency<br />
� R<br />
� vacuum<br />
Resonant frequency<br />
Resonant frequency in vacuum<br />
� Correction factor<br />
References<br />
[1] G. Binnig, C.F. Quate, Ch. Berger, Atomic force microscope, Phys. Rev. Lett. 56 (9)<br />
(1986) 930–933.<br />
[2] W.W. Scott, B. Bhushan, Use of phase imaging in atomic force microscopy for measurement<br />
of viscoelastic contrast in polymer nanocomposites <strong>and</strong> molecularly thick<br />
lubricant films, Ultramicroscopy 97 (2003) 151–169.<br />
[3] P.K. Hansma, J.P. Clevel<strong>and</strong>, M. Radmacher, D.A. Walters, P.E. Hillner, M. Bezanilla,<br />
M. Fritz, D. Vie, H.G. Hansma, C.B. Prater, J. Massie, L. Fukunaga, J. Gurley, V. Elings,<br />
Tapping mode atomic force microscopy in liquids, Appl. Phys. Lett. 64 (13) (1994)<br />
1738–1740.<br />
[4] D. Dowson, C.M. Taylor, Cavitation in bearings, Ann. Rev. Fluid Mech. 11 (1979)<br />
35–66.<br />
[5] Y.H. Zang, J.S. Aspler, M.Y. Boluk, J.H. De Grace, Direct measurement of tensile stress<br />
(“tack”) in thin ink films, J. Rheol. 35 (1991) 345–361.<br />
[6] A. Unsworth, D. Dowson, V. Wright, Cracking joints. A bioengineering study of cavitation<br />
in the metacarpophalangeal joint, Ann. Rheum. Dis. 30 (1971) 348–358.<br />
[7] H. Spikes, The borderline of elastohydrodynamic <strong>and</strong> boundary lubrication, Proc.<br />
Instn. Mech. Eng. 214 (2000) 23–37.<br />
[8] M.T. Tyree, The cohesion-tension theory of sap ascent, J. Exp. Bot. 48 (1997) 1753–1765.<br />
[9] H. Lakrout, P. Sergot, C. Creton, Direct observations of cavitation <strong>and</strong> fibrillation in<br />
a probe tack experiment on model acrylic pressure sensitive adhesive, J. Adhes. 69<br />
(1999) 307–359.<br />
[10] G.J.C. Braithwaite, G.H. McKinley, Microrheometry for studying the rheology <strong>and</strong><br />
dynamics of polymers near interfaces, J. Appl. Rheol. 9 (1999) 27–33.<br />
[11] A. Zosel, The effect of fibrillation on the tack of pressure sensitive adhesives, Int.<br />
J. Adhes. Adhes. 18 (1998) 265–271.<br />
[12] H. Strasburger, Tacky adhesion, J. Colloid Sci. 13 (1958) 218–231.<br />
[13] W.H. Banks, C.C. Mill, Tacky adhesion – a preliminary study, J. Colloid Sci. 8 (1953)<br />
137–147.<br />
[14] F.C. MacKintosh, C.F. Schmidt, Microrheology, Curr. Opin. Coll. Int. Sci. 4 (1999)<br />
300–307.
REFERENCES 271<br />
[15] C. Clasen, G.H. McKinley, Gap-dependent microrheometry of complex liquids,<br />
J. Non-Newtonian Fluid Mech. 124 (2004) 1–10.<br />
[16] A. Mukhopadhyay, S. Granick, Micro- <strong>and</strong> nanorheology, Curr. Opin. Coll. Int. Sci. 6<br />
(2001) 423–429.<br />
[17] P. Attard, Measurement <strong>and</strong> interpretation of elastic <strong>and</strong> viscoelastic properties with<br />
the atomic force microscope, J. Phys. Condens. Matter 19 (2007) 1–33.<br />
[18] A. Maali, B. Bhushan, Nanorheology <strong>and</strong> boundary slip in confined liquids using<br />
atomic force microscopy, J. Phys. Condens. Matter 20 (2008) 1–11.<br />
[19] J.N. Israelachvili, G.E. Adams, Measurement of the forces between two mica surfaces<br />
in aqueous solutions in the range 0–100 nm, J. Chem. Soc. Faraday Trans. 74 (1978)<br />
975–980.<br />
[20] E. Pelletier, J.P. Montfort, F. Lapique, Surface force apparatus <strong>and</strong> its application to<br />
nanorheological studies, J. Rheol. 38 (4) (1994) 1151–1168.<br />
[21] C.D.F. Honig, W.A. Ducker, Squeeze film lubrication in silicone oil: experimental test<br />
of the no-slip boundary condition at solid–liquid interfaces, J. Phys. Chem. C 112<br />
(2008) 17324–17330.<br />
[22] M. Hennemeyer, S. Burghhardt, R.W. Stark, Cantilever micro-rheometer for the characterization<br />
of sugar solutions, Sensors 8 (2008) 10–22.<br />
[23] N. Ahmed, D.F. Nino, V.T. Moy, Measurement of solution viscosity by atomic force<br />
microscopy, Rev. Sci. Instrum. 72 (2001) 2731–2734.<br />
[24] N. Belmiloud, I. Dufour, A. Colin, L. Nicu, Rheological behavior probed by vibrating<br />
microcantilevers, App. Phys. Lett. 92 (2008) 41907.<br />
[25] G.Y. Chen, T. Thundat, E.A. Wachter, R.J. Warmack, Adsorption-induced surface stress<br />
<strong>and</strong> its effects on resonance frequency of microcantilevers, J. Appl. Phys. 77 (8) (1995)<br />
3618–3622.<br />
[26] N.A. Burnham, R.J. Colton, Measuring the nanomechanical properties <strong>and</strong> surface<br />
forces of materials using an atomic force microscope, J. Vac. Sci. Technol. 7 (4) (1989)<br />
2906–2913.<br />
[27] B. Du, O.K.C. Tsui, Q. Zhang, T. He, Study of elastic modulus <strong>and</strong> yield strength of<br />
polymer thin films using atomic force microscopy, Langmuir 17 (2001) 3286–3291.<br />
[28] J. Domke, M. Radmacher, Measuring the elastic properties of thin polymer films with<br />
the atomic force microscope, Langmuir 14 (1998) 3320–3325.<br />
[29] W.A. Ducker, T.J. Senden, R.M. Pashley, Direct measurement of colloidal forces using<br />
an atomic force microscope, Nature 353 (1991) 239–241.<br />
[30] O.I. Vinogradova, G.E. Yakubov, Dynamic effects on force measurements. 2. Lubrication<br />
<strong>and</strong> the atomic force microscope, Langmuir 19 (2003) 1227–1234.<br />
[31] P.G. Hartley, F. Grieser, P. Mulvaney, G.W. Stevens, Surface forces <strong>and</strong> deformation at the<br />
oil–water interface probed using AFM force measurement, Langmuir 15 (1999) 7282–7289.<br />
[32] R.E. Mahaffy, C.F. Shih, F.C. MacKintosh, J. Käs, Scanning probe-based frequencydependent<br />
microrheology of polymer gels <strong>and</strong> biological cells, Phys. Rev. Lett. 85 (4)<br />
(2000) 880–883.<br />
[33] J.L. Alonso, W.H. Goldmann, Feeling the forces: atomic force microscopy in cell biology,<br />
Life Sci. 72 (2003) 2553–2560.<br />
[34] K.E. Bremmell, A. Evans, C.A. Prestidge, Deformation <strong>and</strong> nano-rheology of red<br />
blood cells: an AFM investigation, Coll. Surf. B Biointerfaces 50 (2006) 43–48.<br />
[35] Y-Z. Hu, S. Granick, Microscopic study of thin film lubrication <strong>and</strong> its contributions to<br />
macroscopic tribology, Tribology Lett. 5 (1998) 81–88.<br />
[36] G. Gillies, C.A. Prestidge, P. Attard, An AFM study of the deformation <strong>and</strong> nanorheology<br />
of cross-linked PDMS droplets, Langmuir 18 (2002) 1674–1679.<br />
[37] G. Gillies, C.A. Prestidge, Interaction forces, deformation <strong>and</strong> nano-rheology of emulsion<br />
droplets as determined by colloid probe AFM, Adv. Colloid Interface Sci. 108–109<br />
(2004) 197–205.
272 9. APPLICATION OF ATOMIC FORCE MICROSCOPy<br />
[38] P.A. O’Connell, G.B. McKenna, Rheological measurements of the thermoviscoelastic<br />
response of ultrathin polymer films, Science 307 (2005) 1760–1763.<br />
[39] J.H. Teichroeb, J.A. Forrest, Direct imaging of nanoparticle embedding to probe<br />
viscoelasticity of polymer surfaces, Phys. Rev. Lett. 91 (1) (2003) 16104.<br />
[40] C.M. Mate, G.M. McClell<strong>and</strong>, R. Erl<strong>and</strong>sson, S. Chiang, Atomic-scale friction of a<br />
tungsten tip on a graphite surface, Phys. Rev. Lett. 59 (17) (1987) 1942–1945.<br />
[41] G. Meyer, N.M. Amer, Simultaneous measurement of lateral <strong>and</strong> normal forces with<br />
an optical-beam-deflection atomic force microscope, Appl. Phys. Lett. 57 (20) (1990)<br />
2089–2091.<br />
[42] P.-E. Mazeran, J.-L. Loubet, Normal <strong>and</strong> lateral modulation with a scanning force<br />
microscope, an analysis: implication in quantitative elastic <strong>and</strong> friction imaging,<br />
Tribology Lett. 7 (1999) 199–212.<br />
[43] M. Reinstädtler, U. Rabe, A. Goldade, B. Bhushan, W. Arnold, Investigating ultrathin<br />
lubricant layers using resonant friction force microscopy, Tribology Int. 38 (2005)<br />
533–541.<br />
[44] D.F. Ogletree, R.W. Carpick, M. Salmeron, Calibration of frictional forces in atomic<br />
force microscopy, Rev. Sci. Instrum. 67 (9) (1996) 3298–3306.<br />
[45] S. Moon, M.D. Foster, Near surface nanomechanical behavior of pressure-sensitive<br />
adhesives using lateral modulation scanning probe microscopy, Langmuir 18 (2002)<br />
1865–1871.<br />
[46] A. Paiva, N. Sheller, M.D. Foster, A.J. Crosby, K.R. Shull, Study of the surface adhesion<br />
of pressure-sensitive adhesives by atomic force microscopy <strong>and</strong> spherical indenter<br />
tests, Macromolecules 33 (2000) 1878–1881.<br />
[47] J.P. Phillips, X. Deng, R.R. Stephen, E.L. Fortenberry, M.L. Todd, D.M. McClusky,<br />
S. Stevenson, R. Misra, S. Morgan, T.E. Long, Nano- <strong>and</strong> bulk-tack adhesive properties<br />
of stimuli-responsive fullerene-polymer blends, containing polystyrene-block-<br />
polybutadiene-block-polystyrene <strong>and</strong> polystyrene-block-polyisoprene-block-polystyrene<br />
rubber-based adhesives, Polymer 48 (2007) 6773–6781.<br />
[48] J.W.G. Tyrrell, P. Attard, Viscoelastic study using an atomic force microscope modified<br />
to operate as a nanorheometer, Langmuir 19 (2003) 5254–5260.<br />
[49] C. Cottin-Bizonne, S. Jurine, J. Baudry, J. Crassous, F. Restagno, E. Charlaix,<br />
Nanorheology: an investigation of the boundary condition at hydrophobic <strong>and</strong><br />
hydrophilic interfaces, Eur. Phys. J. E 9 (2002) 47–53.<br />
[50] R. García, R. Pérez, Dynamic atomic force microscopy methods, Surf. Sci. Rep. 47<br />
(2002) 197–301.<br />
[51] N.A. Burnham, G. Gremaud, A.J. Kulik, P.-J. Gallo, F. Oulevey, Materials properties<br />
measurements: choosing the optimal scanning probe microscope configuration, J. Vac.<br />
Sci. Technol. B 14 (2) (1996) 1308–1312.<br />
[52] P. Maivald, H.J. Butt, S.A.C. Gould, C.B. Prater, B. Drake, J.A. Gurley, V.B. Elings,<br />
P.K. Hansma, Using force modulation to image surface elasticities with the atomic<br />
force microscope, Nanotechnology 2 (1991) 103–106.<br />
[53] A.A. Galuska, R.R. Poulter, K.O McElrath, Force modulation AFM of elastomer blends:<br />
morphology, fillers <strong>and</strong> cross-linking, Surface Interface Anal. 25 (1997) 418–429.<br />
[54] B. Gotsmann, C. Seidel, B. Anczykowski, H. Fuchs, Conservative <strong>and</strong> dissipative<br />
tip–sample interaction forces probed with dynamic AFM, Phys. Rev. B 60 (15) (1999)<br />
11051–11061.<br />
[55] H. Watabe, K. Nakajima, Y. Sakai, T. Nishi, Dynamic force spectroscopy on a single<br />
polymer chain, Macromolecules 39 (2006) 5921–5925.<br />
[56] S.M. Notley, S. Biggs, V.S.J. Craig, Application of a dynamic atomic force microscope<br />
for the measurement of lubrication forces <strong>and</strong> hydrodynamic thickness between surfaces<br />
bearing adsorbed polyelectrolyte layers, Macromolecules 36 (2003) 2903–2906.
REFERENCES 273<br />
[57] R.M. Overney, D.P. Leta, C.F. Pictroski, M.H. Rafailovich, Y. Liu, J. Quinn, J. Sokolov,<br />
A. Eisenberg, G. Overney, Compliance measurements of confined polystyrene solutions<br />
by atomic force microscopy, Phys. Rev. Lett. 76 (8) (1996) 1272–1275.<br />
[58] F. Oulevey, N.A. Burnham, G. Gremaud, A.J. Kulik, H.M. Pollock, A. Hammiche,<br />
M. Reading, M. Song, D.J. Hourston, Dynamic mechanical analysis at the submicron<br />
scale, Polymer 41 (2000) 3087–3092.<br />
[59] J.B. Pethica, W.C. Oliver, Tip surface interactions in STM <strong>and</strong> AFM, Physica Scripta<br />
T19 (1987) 61–66.<br />
[60] U. Rabe, S. Amelio, E. Kester, V. Scherer, S. Hirsekorn, W. Arnold, Quantitative determination<br />
of contact stiffness using atomic force acoustic microscopy, Ultrasonics 38<br />
(2000) 430–437.<br />
[61] D.C. Hurley, K. Shen, N.M. Jennett, J.A. Turner, Atomic force acoustic microscopy methods<br />
to determine thin-film elastic properties, J. Appl. Phys. 94 (4) (2003) 2347–2354.<br />
[62] A.R. Suraya, P.F. Luckham, C.J. Lawrence, Shear thinning <strong>and</strong> frequency dependent<br />
behaviour of adsorbed polymer layers: Part I. Experimental aspects <strong>and</strong> a first order<br />
analysis, J. Non-Newtonian Fluid Mech. 148 (2008) 57–64.<br />
[63] M.C. Friedenberg, C.M. Mate, Dynamic viscoelastic properties of liquid polymer films<br />
studied by atomic force microscopy, Langmuir 12 (1996) 6138–6142.<br />
[64] J. Choi, T. Kato, Measurement on the shear properties of liquid nanomeniscus bridge,<br />
Tribology Int. 36 (2003) 475–481.<br />
[65] G.J.C. Braithwaite, P.F. Luckham, The simultaneous determination of the forces <strong>and</strong><br />
viscoelastic properties of adsorbed polymer layers, J. Coll. Int. Sci. 218 (1999) 97–111.<br />
[66] M. Radmacher, R.W. Tillmann, H.E. Gaub, Imaging viscoelasticity by force modulation<br />
with the atomic force microscope, Biophys. J. 64 (1993) 735–742.<br />
[67] P.I. Oden, G.Y. Chen, R.A. Steele, R.J. Warmack, T. Thundat, Viscous drag measurements<br />
utilizing microfabricated cantilevers, Appl. Phys. Lett. 68 (26) (1996) 3814–3816.<br />
[68] J.E. Sader, Frequency response of cantilever beams immersed in viscous fluids with<br />
applications to the atomic force microscope, J. Appl. Phys. 84 (1998) 64–76.<br />
[69] E.O. Tuck, J. Eng. Math. 3 (1969) 29.<br />
[70] S. Boskovic, J.W.M. Chon, P. Mulvaney, J.E. Sader, Rheological measurements using<br />
microcantilevers, J. Rheol. 46 (4) (2002) 891–899.<br />
[71] A. Maali, C. Hurth, R. Boisgard, C. Jai, T. Cohen-Bouhacina, J.-P. Aime, Hydrodynamics<br />
of oscillating atomic force microscopy cantilevers in viscous fluids, J. Appl. Phys.<br />
97 (2005) 74907.<br />
[72] J.W.M. Chon, P. Mulvaney, J.E. Sader, Experimental validation of theoretical models<br />
for the frequency response of atomic force microscope cantilever beams immersed in<br />
fluids, J. App. Phys. 87 (8) (2000) 3978–3988.<br />
[73] M. Gelbert, M. Biesalski, J. Rühe, D. Johannsmann, Collapse of polyelectrolyte brushes<br />
probed by noise analysis of a scanning force microscope cantilever, Langmuir 16<br />
(2000) 5774–5784.<br />
[74] F. Benmouna, D. Johannsmann, Hydrodynamic interaction of AFM cantilevers with<br />
solid walls: an investigation based on AFM noise analysis, Eur. Phys. J. 9 (2002) 435–441.<br />
[75] F. Benmouna, D. Johannsmann, Viscoelasticity of gelatin surfaces probed by AFM<br />
noise analysis, Langmuir 20 (2004) 188–193.<br />
[76] A. Ouibrahim, D.H. Fruman, R. Gaudemer, Vapor cavitation in very confined spaces<br />
for Newtonian <strong>and</strong> non Newtonian fluids, Phys. Fluids 8 (1996) 1964–1971.<br />
[77] T. Narumi, T. Hasegawa, Experimental study on the squeezing flow of viscoelastic<br />
fluids. 1. The effect of liquid properties on the flow between a spherical surface <strong>and</strong> a<br />
flat plate, Bull. JSME 29 (1986) 3731–3736.<br />
[78] D.D. Joseph, Cavitation <strong>and</strong> the state of stress in a flowing liquid, J. Fluid Mech. 366<br />
(1998) 367–378.
274 9. APPLICATION OF ATOMIC FORCE MICROSCOPy<br />
[79] K.H. Christenson, P.M. Claesson, Cavitation <strong>and</strong> the interaction between macroscopic<br />
hydrophobic surfaces, Science 239 (1988) 390–392.<br />
[80] J.C. Fisher, The fracture of liquids, J. Appl. Phys. 19 (1948) 1062–1067.<br />
[81] R.J. Speedy, Stability-limit conjecture. An interpretation of the properties of water,<br />
J. Phys. Chem. 86 (1982) 982–991.<br />
[82] Q. Zheng, D.J. Durben, G.H. Wolf, C.A. Angell, Liquids at large negative pressure:<br />
water at the homogeneous nucleation limit, Science 254 (1991) 829–832.<br />
[83] D.R. Berard, P. Attard, G.N. Patey, Cavitation of a Lennard–Jones fluid between hard<br />
walls, <strong>and</strong> the possible relevance to the attraction measured between hydrophobic<br />
surfaces, J. Chem. Phys. 98 (1993) 7236–7244.<br />
[84] P. Attard, Bridging bubbles between hydrophobic surfaces, Langmuir 12 (1996)<br />
1693–1695.<br />
[85] Y.L. Chen, T. Kuhl, J. Israelachvili, Mechanism of cavitation damage in thin liquid<br />
films: collapse damage vs. inception damage, Wear 153 (1992) 31–51.<br />
[86] T. Kuhl, M. Ruths, Y.L. Chen, J. Israelachvili, Direct visualisation of cavitation <strong>and</strong><br />
damage in ultrathin liquid films, J. Heart Valve Dis. 3 (1994) 117–127.<br />
[87] L. Rayleigh, On the pressure developed in a liquid during the collapse of a spherical<br />
cavity, Phil. Mag. 34 (1917) 94–98.<br />
[88] M.S. Plessett, A. Prosperetti, Bubble dynamics <strong>and</strong> cavitation, Ann. Rev. Fluid Mech. 9<br />
(1977) 145–185.<br />
[89] J.R. Blake, D.C. Gibson, Growth <strong>and</strong> collapse of a vapour cavity near a free surface,<br />
J. Fluid Mech. I 11 (1981) 123–140.<br />
[90] T.B. Benjamin, A.T. Ellis, The collapse of cavitation bubbles <strong>and</strong> the pressures thereby<br />
produced against solid boundaries, Phil. Trans. R. Soc. A 260 (1966) 221–240.<br />
[91] T.G. von Leeuwen, J.H. Meertens, E. Velema, M.J. Post, C. Borst, Intraluminal vapor<br />
bubble induced by excimer laser pulse causes microsecond arterial dilation <strong>and</strong><br />
invagination leading to extensive wall damage in the rabbit, Circulation 87 (1993)<br />
1258–1263.<br />
[92] R. Vorreuther, R. Corleis, T. Klotz, P. Bernards, U. Engelmann, Impact of shock wave<br />
pattern <strong>and</strong> cavitation bubble size on tissue damage during ureteroscopic electrohydraulic<br />
lithotripsy, J. Urol. (Baltimore) 153 (1995) 849–853.<br />
[93] P. Zhong, I. Cionta, S. Zhu, F.H. Cocks, G.M. Preminger, Effects of tissue constraint<br />
on shock wave-induced bubble expansion in vivo, J. Acoust. Soc. Am. 104 (1998)<br />
3126–3129.
C H A P T E R<br />
10<br />
Future Prospects<br />
The preceding chapters have shown that the use of AFM has already<br />
made substantial contributions to underst<strong>and</strong>ing <strong>and</strong> improving processes<br />
across a wide range of engineering. The literature on such applications is<br />
very extensive, so the aim of the present book has been to focus on some<br />
key examples in depth rather than to attempt a comprehensive overview,<br />
which would necessarily have been either unreasonably lengthy or else<br />
unsatisfyingly superficial. It is our hope that we have given an account of<br />
the principles, achievements <strong>and</strong> possibilities of AFM that is sufficiently<br />
informative for experts in other areas of process engineering to envisage<br />
how their own work may be enhanced using the techniques described.<br />
Owing to the broadness of the field, it is difficult to envisage where the<br />
most important developments are likely to take place in the future. However,<br />
we have asked the authors of the preceding chapters to give a brief<br />
account as to how they would see the use of AFM in their own specialties<br />
developing, with the following results.<br />
The importance of AFM lies in its capability to provide better underst<strong>and</strong>ing<br />
of materials’ structure, surface characteristics <strong>and</strong> the interactive<br />
forces at the meso- <strong>and</strong> nanoscale level. This will greatly help underst<strong>and</strong><br />
large-scale engineering processes, especially as materials are increasingly<br />
being designed down to the submicrometre level. The developments of<br />
colloid, coated-colloid <strong>and</strong> cell probe techniques have already opened<br />
new windows of applications to a variety of engineering processes including<br />
those involving bubbles, such as flotation, <strong>and</strong> processes within the<br />
biotechnology sector. For example, characterisation of membrane surface<br />
morphology <strong>and</strong> forces of interaction with colloidal particles <strong>and</strong> cells<br />
has facilitated the development of synthetic membranes with greatly<br />
improved process-separation characteristics. As synthetic membranes play<br />
a key role in water treatment, including desalination, such use of AFM is<br />
likely to be an increasingly important application. Similarly, AFM studies<br />
of particle-bubble interactions can play a major role in optimising mineral<br />
separations, one of the largest-scale of all process operations. However,<br />
Atomic Force Microscopy in Process Engineering 275<br />
© 2009, Elsevier Ltd
276 10. FuTuRE PRosPECTs<br />
most AFM studies have so far been carried out under ambient conditions,<br />
whereas many industrial processes operate at reduced or elevated temperatures<br />
<strong>and</strong> pressures. The more recent development of hot–cold stages<br />
has made possible the application of AFM for both imaging <strong>and</strong> quantification<br />
of forces of interactions in the temperature range between �90°C<br />
<strong>and</strong> 250°C. Future development of pressure cells to enable the operation<br />
at both vacuum <strong>and</strong> pressures higher than atmospheric will further widen<br />
the range of process-relevant environments for further studies.<br />
The application of process engineering knowledge to biotechnology<br />
<strong>and</strong> medicine is showing huge potential. The integration of cutting-edge<br />
techniques to develop tailored cellular micro-environments as model<br />
systems has proven essential for the systematic investigation of a number<br />
of physiological processes in cell biology. From these studies, it has<br />
become apparent that restoring native functionalities using smart extracellular<br />
matrix (ECM), or tissue, depends on adequate consideration of<br />
interconnected processes that are regulated by biochemical, physical <strong>and</strong><br />
mechanical factors in the ECM.<br />
AFM, with its proven capability for nanoscale measurements of biomolecular<br />
interactions, <strong>and</strong> physical <strong>and</strong> mechanical properties of materials,<br />
is expected to continue to make invaluable contributions to cell biology<br />
<strong>and</strong> biotechnology fields. However, new developments are desirable<br />
to improve measurement reliability <strong>and</strong> to underst<strong>and</strong> the factors that<br />
determine measurement reproducibility. In the same way as closed-loop<br />
scanners have greatly enhanced the metrological capabilities of AFM <strong>and</strong><br />
hence the precision in measurement of large biological samples (such as<br />
living cells), there is still much scope for improvements associated with<br />
force measurements. This may be in the area of probe fabrication, providing<br />
cheap sources of cantilevers with high-tolerance spring constants or more<br />
likely in reliable, accurate, non-destructive <strong>and</strong> (hopefully) simple protocols<br />
for the calibration of existing probe types. Such improvements may<br />
in turn provide an impetus for the development <strong>and</strong> widespread use of<br />
mathematical modelling (such as the use of finite element techniques) to<br />
interpret force curves in the context of heterogeneous cellular structures –<br />
modern computing now makes such data fitting a tractable problem. Such<br />
approaches require close collaboration between biologists, engineers <strong>and</strong><br />
physical scientists.<br />
The integration of AFM <strong>and</strong> optical techniques for simultaneous interrogation<br />
of biochemical functionalities with physical/mechanical properties<br />
of a cell offers huge benefits <strong>and</strong> is likely to attract increasing research<br />
efforts. High-speed optical imaging offers the possibility of monitoring<br />
processes that are currently too fast for AFM to interrogate. The present<br />
developments of high-speed AFM imaging are certainly encouraging.<br />
However, progress is required to minimise the tip-surface interactions<br />
that, at these speeds, greatly perturb the surface being imaged.
10. FuTuRE PRosPECTs 277<br />
In recent years, a number of complementary measurement probes have<br />
been integrated onto AFM tips, permitting the imaging of other properties<br />
of a sample simultaneously with its topography. These novel developments<br />
have included the incorporation of electrochemical, thermal, magnetic <strong>and</strong><br />
electronic sensors. Researchers have also found that, in addition to being<br />
able to perform measurement functions, often these sensors allow the<br />
controlled alteration of the local environment around the tip. Preliminary<br />
work using certain types of sensors has shown that this feature could find<br />
great applicability in modulating micro- <strong>and</strong> nanoenvironments of living<br />
cells. Such functionalised probes could be incorporated into an existing<br />
AFM system, hence significantly enhancing its capability at modest cost.<br />
The availability of high-quality probes is also a factor limiting the uptake<br />
of some combined techniques, such as AFM <strong>and</strong> scanning near field optical<br />
microscopy (SNOM). The combination of super-resolution optical <strong>and</strong><br />
topographic imaging (AFM-SNOM) is certainly of great interest to biologists<br />
because of the wealth of new information it could provide.<br />
Polymers on surfaces play <strong>and</strong> will continue to play a major role in various<br />
engineering applications such as colloidal stabilisation, lab-on-a-chip<br />
devices, polymer composites <strong>and</strong> nanocomposites. Surprisingly, the significance<br />
of the polymer-solid interface region has been realised only relatively<br />
recently, <strong>and</strong> there are many gaps in our basic underst<strong>and</strong>ing. In nanocomposites,<br />
the interface regions can be so extensive that they can occupy most<br />
of the overall volume of the material even at low or moderate nanoparticle<br />
loadings. Furthermore, the conformation <strong>and</strong> nature of entanglements of<br />
confined polymer chains in close proximity to a solid surface can deviate<br />
significantly from the bulk. As AFM is a real-space, high-resolution surface<br />
technique capable of probing quantitatively both the nanostructural <strong>and</strong><br />
nanomechanical properties, it is likely to play an important role in the study<br />
of the interface region in polymer composites <strong>and</strong> nanocomposites.<br />
The self-assembly <strong>and</strong> self-organisation of polymers on solid–liquid<br />
interfaces can be used as a generic technique for the inexpensive mass<br />
production of surface nanostructures <strong>and</strong> nanopatterns. Many issues<br />
remain to be elucidated by careful quantitative experimentation, <strong>and</strong> AFM<br />
is being proven to be an indispensable tool for the study of such systems<br />
in air or in liquids. Furthermore, the recent development of high-speed<br />
AFM with image-acquisition times in the order of a few milliseconds is an<br />
important development that could allow the monitoring of the mobility<br />
of polymer chains on a solid surface <strong>and</strong> consequently the kinetic paths<br />
of the nanostructures/nanopatterns formation. AFM offers the possibility<br />
for systematic studies of such systems that can elucidate the underlying<br />
physical mechanisms of the processes occurring on solid surfaces. These<br />
studies can lead in the construction of phase/state diagrams, which may<br />
be used for the design of well-controlled <strong>and</strong> specified surface nanostructures<br />
for applications spanning from nanoelectronic devices to vectors for
278 10. FuTuRE PRosPECTs<br />
targeted drug/gene delivery. Further, AFM is becoming an indispensable<br />
characterisation tool of miniaturised device components, which increasingly<br />
involve the use of (bio)polymers, ultrathin films <strong>and</strong> nanostructures<br />
on surfaces.<br />
To date, the successful use of AFM-based techniques to determine the<br />
viscoelastic properties of materials is exemplified by the successful use of<br />
indentation or compression experiments, for which established theoretical<br />
models adequately describe the rheological properties. Of particular<br />
note are the developments in biomechanical studies, wherein the determination<br />
of the viscoelastic response of cells <strong>and</strong> biological tissues using a<br />
combination of an AFM <strong>and</strong> a confocal laser scanning microscope (CLSM)<br />
is an interesting development, especially so with the use of modern fastscanning<br />
CLSMs.<br />
For fluids, although experimental studies have shown that an AFM is<br />
certainly capable of emulating the rheological capabilities of the surface<br />
forces apparatus (SFA), the use of an AFM to quantitatively describe the<br />
rheology of thin fluid films remains the subject of considerable research<br />
interest. Although, the viability of AFM as a micro-rheometrical tool<br />
has been established, the development of mathematical models capable<br />
of interpreting the response of microcantilevers in viscoelastic fluids is<br />
still evolving, <strong>and</strong> further development is vital to the successful application<br />
of this technique. Nevertheless, the integration of AFM cantilevers in<br />
microfluidic <strong>and</strong> other fluid sensor devices is a realistic option, enabled<br />
most significantly by advances in the description of viscous effects upon<br />
resonance characteristics. From a rheometrical perspective, this is particularly<br />
encouraging as few other technologies can address the associated<br />
issues of scale. The development of this area will in part rely upon the<br />
fabrication of appropriate cantilever geometries, such that the resonance<br />
characteristics <strong>and</strong> sensitivity of the device are optimized for the fluid of<br />
interest. In this respect, the fabrication of multilevers on common substrates<br />
may extend the measuring capabilities <strong>and</strong> represents a potential<br />
enabling technology for the study of more complex, multicomponent fluidic<br />
systems.<br />
The adaptation of dynamic AFM methods for the routine rheological<br />
characterisation of thin fluid films is more problematical. However,<br />
qualitative results suggest that the technique has potential merits. The<br />
dynamic response of a vibrating probe is clearly able to detect the onset<br />
of viscoelastic behaviour, but quantitative rheological information as yet<br />
remains elusive. Surprisingly, many dynamic studies favour oscillation<br />
in the direction normal to the surface, <strong>and</strong> as such it may be anticipated<br />
that significant compressional wave components can influence oscillation<br />
of the probe. For the analogous study of macroscale oscillatory shear<br />
<strong>and</strong> microscale rheometrical parameters, further studies on oscillatory<br />
microscale shear deformations would be beneficial.
10. FuTuRE PRosPECTs 279<br />
In either case, as the system response is not simply related to the stress<br />
<strong>and</strong> the strain in the fluid, there is at present no satisfactory way to quantitatively<br />
determine the complex moduli. However, the promising results<br />
of studies to date suggest that an AFM operating in dynamic mode is<br />
suited to adaptation as a mechanical interferometer, i.e. a device that can<br />
be used to determine the ‘time of flight’ of small amplitude shear waves<br />
across a fluid-filled gap. This is analogous to the so-called virtual gap<br />
rheometer (VGR), wherein the phase delays determined at two different<br />
locations in the fluid can be used to determine the phase velocity of the<br />
shear wave. By studying several frequencies, the frequency-dependent<br />
behaviour of viscoelastic systems may be exploited to determine the gel<br />
point of curing polymers.<br />
Most importantly, the future prospects of all applications of AFM in<br />
process engineering will be greatly enhanced by increased collaboration<br />
between engineers, physical scientists, biological scientists, mathematicians<br />
<strong>and</strong> instrument manufacturers. Among the most crucial of the multidisciplinary<br />
themes that need to be addressed are the development of probes of<br />
highly specified <strong>and</strong> reproducible physical properties; the development of<br />
high-speed image acquisition so that dynamic processes may be followed;<br />
the combination of AFM with other techniques in a single instrument,<br />
especially techniques that allow chemical characterisation of surfaces <strong>and</strong><br />
the integration of AFM data acquisition <strong>and</strong> advanced mathematical<br />
modelling software for data interpretation.
A<br />
Actin stress fiber, 197<br />
Adhesion, 67–70, 198, 202, 226, 234, 246<br />
drug particles, 174–76, 178–79, 182, 184<br />
forces, 67–70, 81, 91–93, 97, 141–43,<br />
151–63, 164–66<br />
particles to bubbles, 83–85, 90, 91<br />
particles to membranes, 112, 113, 122<br />
Adhesives, 248–49<br />
AFM-optical microscope, 211<br />
Aggregation, 229, 236<br />
Apparatus coefficient, 255<br />
Arg-Gly-Asp (RGD), 197, 203–04<br />
Atomic force acoustic microscopy, 252<br />
B<br />
Biaxial extension, 248<br />
Biocompatibility, 226<br />
Biofilms, 135<br />
Block copolymer, 228, 236–237<br />
Boundary conditions, 49<br />
charge regulation, 49<br />
constant surface charge, 49<br />
constant surface potential, 49<br />
constant zeta potential, 49<br />
Bovine Serum Albumin (BSA), 49<br />
C<br />
Capillary forces, 10, 15, 18, 90, 94, 230<br />
Cauchy Plot, 41<br />
Cavitation, 246–47, 259–61<br />
Cell membrane, 197<br />
Collagen fibres, 205<br />
Cell model, 52<br />
Cell probe, 113<br />
Charging of surface, 44<br />
Chemisorption, 227<br />
Chloride ion binding, 51<br />
Coating flows, 247<br />
Colloid probe, 22–23, 32, 55, 69, 111, 112,<br />
210, 248, 253–55, 265–67<br />
coated, 113<br />
membrane measurements, 151–63, 164–66<br />
particle bubble measurements, 82, 91–93,<br />
96, 99<br />
Index<br />
281<br />
Colloidal gas aphrons, 95–97<br />
Colloidal lithography, 203, 207<br />
Cytoplasm, 197<br />
Cytoskeleton, 197, 212, 217<br />
Complex modulus, 252–53<br />
Confined fluids, 252–56, 259–67<br />
Contact angles, 88–91<br />
Contact mode, 9<br />
Convolution, 230–31<br />
Corrosion metal, 132<br />
Counterions, 47<br />
Critical flux, 123<br />
D<br />
Damping coefficient, 254–55<br />
Density, 247, 256–59<br />
Derjaguin approximation, 33<br />
Derjaguin, Muller Toporov model,<br />
67, 178<br />
Dewetting, 229, 236<br />
DLVO theory, 53–54, 154<br />
forces, 54–58, 91, 93<br />
DMT model, see Derjaguin, Muller Toporov<br />
model<br />
Dry powder inhalator, 174–76<br />
Dynamic AFM, 2, 10–11, 249–56<br />
E<br />
Effective stiffness, 43<br />
Elastic modulus, 180–81<br />
see also Young’s modulus<br />
Elasticity, 214<br />
Electrical double layer interactions, 47–53<br />
analytical solutions, 47<br />
numerical solutions, 48<br />
Electromechanical oscillator, 252<br />
Electron beam lithography (EBL), 203, 208<br />
Extracellular matrix (ECM), 195, 205<br />
proteins, 196, 197, 207<br />
Electropolishing, 130<br />
Electrostatic interactions, 111, 121, 227<br />
Entropic elasticity, 234<br />
Exponential damping length, 258<br />
Extensional flow, 261<br />
Extensional rheometer, 261
282 InDEx<br />
F<br />
Filamentation, 259, 261–62, 264–67<br />
Film splitting, 247, 267<br />
Filopodia, 198, 207, 212<br />
Focal adhesion, 197–8<br />
Focal complex, 198<br />
Force between two spheres, 33<br />
Force distance curve, 111, 112, 117<br />
Force map, 250<br />
Force modulation, 12, 249<br />
Force oscillations, 61<br />
Force volume imaging, 11<br />
Formulation, 174–75, 178, 182, 185–87,<br />
188–90<br />
Fourier transform, 256<br />
Friction force microscopy, 21, 248<br />
Frictional force mode, 12<br />
Frictional forces, 12, 63, 178–79, 248<br />
see also Lateral force<br />
Froth flotation, 82<br />
Functionalised surface, 199<br />
G<br />
Gouy-Chapman theory, 45–46<br />
Grafting density, 228, 231–233<br />
H<br />
Hamaker constant, 38–42<br />
calculation of, 40<br />
pairwise addition, 38<br />
Hardness, 180<br />
Hertz model, 12, 180–82, 214–7, 248<br />
High-speed video microscopy, 263<br />
Humic acid, 166<br />
Humic substances, 140<br />
Humidity, 175–77, 183, 186<br />
Hydration forces, 59–60<br />
Hydrodynamic drag, 66<br />
Hydrodynamic forces, 63, 84–85, 98–100<br />
Hydrodynamic function, 17, 257–58<br />
Hydrodynamic lubrication, 253–54<br />
Hydrophobic surfaces, 64<br />
I<br />
Imaging in liquid, 116, 123<br />
Indentation, 179–82, 215, 247, 251<br />
Inner Helmholtz plane (IHP), 46<br />
Integrins, 197<br />
Intermittent contact mode, see Tapping<br />
mode<br />
Isopotential lines, 119<br />
J<br />
JKR model, see Johnson, Kendall, Roberts<br />
model<br />
Johnson, Kendall, Roberts model, 12, 67,<br />
84, 178<br />
L<br />
Lamellipodia, 198<br />
Living cells, 212–4<br />
Lateral force, 12–22, 248–49<br />
Lateral resolution, 226<br />
Lennard-Jones potential, 36–37<br />
Lifshitz theory, 38, 40<br />
Light intensifier, 264–65<br />
Line tension, 89<br />
Liquid bridge, 255, 261, 265–68<br />
Liquid structure, 58–59<br />
Loading force, 234–35<br />
bubble, 97<br />
membrane, 151–60<br />
Loss modulus, 253<br />
M<br />
Macromolecules, 225, 229<br />
Membrane characterisation, 141<br />
Membranes, 57, 107–26<br />
development, 124–26<br />
fouling, 57, 68, 140–44<br />
Mesoscale, 246, 261<br />
Metal surfaces, 126<br />
Metastability, 260<br />
Microcontact printing, 201<br />
Micropatterning, 199<br />
Microsphere indentation, 210<br />
Microfilament, 264–67<br />
Microfiltration, 108<br />
Microfoams, 95–97<br />
Micromanipulator, 34<br />
Microrheometry, 247, 261–68<br />
Microsensor, 247<br />
Modification of filtration membranes,<br />
139–68<br />
Molecular weight cut-off, 113<br />
Monolayers, 226–30, 233–34, 236–39<br />
Multiparticle interaction, 52–53<br />
n<br />
Nanofiltration, 108<br />
Nanoimprint lithography (NIL), 203<br />
Nanopatterning, 203, 226–28, 233<br />
Nanotopography, 205–6
Natural organic matter, 140<br />
Negative pressure, 247<br />
Non-contact mode, 10–11<br />
Non-DLVO forces, 58<br />
O<br />
Optical microscope, 262, 264<br />
Outer Helmholtz plane (OHP), 46<br />
P<br />
Particle-bubble interactions, 81–106<br />
Pharmaceuticals, 173–90<br />
Phase angle, 250, 254–56<br />
Phase imaging, 10, 185–86<br />
Photolithography, 199<br />
Physisorption, 199<br />
Polymer demixing, 206<br />
Physisorption, 227, 230, 235<br />
Pinned micelles, 232<br />
Poisson-Boltzmann Equation, 45<br />
Polar liquids, 55<br />
Polymer brush, 228–36<br />
Polymer chains, 226–34, 237–41<br />
Polymer coatings, 225–26<br />
Pore size diameter, 115 distribution, 109,<br />
110, 116, 118, 119<br />
Pore size distribution, 140, 147–50, 168<br />
Powders, 174–75<br />
Primary alcohols, 61<br />
Q<br />
Q factor, see Quality factor<br />
Quality factor, 17<br />
R<br />
Resonance spectra, 256–59<br />
Resonant frequency, 17, 21, 258<br />
Retardation, 40<br />
Reverse osmosis, 108<br />
Reynolds number, 257<br />
Rheology, 246, 252–56, 262<br />
Roughness, 42, 44, 55, 69–70, 115, 120, 128,<br />
129, 134<br />
S<br />
Scanning probe lithography, 203<br />
Scanning thermal microscope, 186–90<br />
Self assembled monolayers (SAM), 199<br />
Self-assembly nanofabrication, 203<br />
SEM, 210<br />
InDEx 283<br />
Shear strain, 252, 255<br />
Shear stress, 252, 255<br />
Simple harmonic oscillator, 257<br />
Soft lithography, 201<br />
Solvation forces, 58–62<br />
Spring constant, 176–77<br />
Spring constant calibration, 16–22<br />
Spring constants, 87<br />
Star polymers, 237–240<br />
Steel, 132<br />
Steric interactions, 62<br />
Storage modulus, 253<br />
Stress fiber, 208<br />
Surface electrical properties, 117<br />
T<br />
Tack, 247, 249<br />
Tapping mode, 10, 236–37, 240, 250<br />
TEM, 210<br />
Tension, 246, 259–60<br />
Theta solvents, 62<br />
Thin liquid films, 254, 260<br />
Three-phase contact, 83, 88–90, 94–95, 97<br />
Threshold force, 94–95, 97<br />
Torsional spring constant, 21–22<br />
U<br />
Ultrafiltration, 108<br />
Uniaxial extension, 267<br />
V<br />
van der Waals forces, 35–44, 83–84,<br />
91, 227<br />
Debye forces, 35–37<br />
Keesom forces, 35–37<br />
London dispersion forces, 35–37<br />
Measurement by AFM, 42<br />
Viscoelasticity, 249, 252–53, 255, 259<br />
Viscosity, 247, 255–59<br />
W<br />
Work of adhesion, 43, 84<br />
Y<br />
Young equation, 88–89<br />
Young’s modulus (E), 180–82, 215, 247<br />
see also Elastic modulus<br />
Z<br />
Zeroth-order Stern model, 51